第 22 章低层视觉处理：视网膜

视网膜是大脑观察世界的窗口。所有的

视觉经验

都基于眼睛中这个神经回路处理信息。视网膜的输出仅通

过一百万个视神经纤维传送到大脑，但几乎一半的大脑皮层用于处理这些信号。由于设计或缺陷而丢失在视网

膜中的视觉信息永远无法恢复。由于视网膜处理对可以看到的内容设置了基本限制，因此人们对了解视网膜的

功能非常感兴趣。

从表面上看，脊椎动物的眼睛看起来很像相机。如图 22.0.1 所示，瞳孔形成可变孔径，角膜和晶状体提供折

射光学器件，将外部世界的小图像投射到眼球后部的光敏视网膜上。但这就是眼睛和相机类比结束的地方。如

图 22.0.2 所示，视网膜是一层薄薄的神经元，厚几百微米，由 5 种主要细胞类型组成，这些细胞排列在由 2 个突

触层隔开的 3 个细胞层中。

注视点

光线

角膜

视网膜

中央凹

视神经

视盘

色素上皮层

晶状体

小凹

光线

光感受器

色素上皮层

双极细胞

神经节细胞

视网膜

A 光线折射到视网膜上 B 光在中央凹的聚焦

光感受器在中央凹的堆积

10 微米

图 22.0.1: 眼睛将视觉场景投射到视网膜的感光器上。A. 来自视野中物体的光被角膜和晶状体折射并聚焦到视网

膜上。B. 在中央凹，对应于凝视的正中心，视网膜的近端神经元被移到一边，因此光线可以直接进入光受体。C.

用于评估正常视力的视力表中的字母被投射到中央凹中密集的光受体上。由于肉眼的光学衍射，虽然不像这里

显示的那样清晰，但字母的最小可辨别笔画宽度约为一个圆锥直径

[177]

。

最外层的感光细胞吸收光并将其转化为神经信号，这一过程称为光转导。这些信号通过突触传递到双极细

胞，双极细胞又连接到最内层的视网膜神经节细胞。视网膜神经节细胞是视网膜的输出神经元，它们的轴突形

成视神经。如图 22.0.3 所示，除了从感觉神经元到输出神经元的这种直接通路外，视网膜回路还包括许多由外

部突触层中的水平细胞和内部突触层中的无长突细胞提供的横向连接。

视网膜回路执行低层视觉处理，这是视觉图像分析的初始阶段。它从眼睛的原始图像中提取某些空间和时

间特征，并将它们传送到更高的视觉中心。这种处理的规则适应环境条件的变化。特别是，视网膜必须调整其对

不断变化的照明条件的敏感度。尽管在每天的过程中遇到的光强度范围很大，但这种适应使我们的视力或多或

少保持稳定。

在本章中，我们依次讨论了视网膜功能的 3 个重要方面：光转导、预处理和适应。我们将说明实现它们的神

经机制及其对视觉感知的影响。

22.1 光感层对视觉图像进行采样

Chapter 22 / Low-Level Visual Processing: The Retina 523

Figure 22–2 The retina comprises five distinct layers of

neurons and synapses.

A.A perpendicular section of the human retina seen through the

light microscope. Three layers of cell bodies are evident. The outer

nuclear layer contains cell bodies of photoreceptors; the inner

nuclear layer includes horizontal, bipolar, and amacrine cells; and

the ganglion cell layer contains ganglion cells and some displaced

amacrine cells. Two layers of fibers and synapses separate these:

the outer plexiform layer and the inner plexiform layer.

(Reproduced, with permission, from Boycott and Dowling 1969.

Permission conveyed through Copyright Clearance Center.)

B.Neurons in the retina of the macaque monkey based on

Golgi staining. The cellular and synaptic layers are aligned with

the image in part A. (Abbreviations: M ganglion, magnocellular

ganglion cell; P ganglion, parvocellular ganglion cell.) (Repro-

duced, with permission, from Polyak 1941.)

外核层

外网状层

内核层

内网状层

神经节

细胞层

无长突细胞

视锥

分散

双极细胞

水平细胞

大

神经节细胞

大

神经节细胞

侏儒

双极细胞

小

神经节细胞

视杆

杆状

双极细胞

A 视网膜切片 B 视网膜中的神经元

50 微米

approaches the limit imposed by diffraction at the

pupil. Away from the axis, the image is degraded sig-

nificantly owing to aberrations in the cornea and lens

and may be degraded further by abnormal conditions

such as light-scattering cataracts or refractive errors

such as myopia.

The area of retina near the optical axis, the fovea, is

where vision is sharpest and corresponds to the center

of gaze that we direct toward the objects of our atten-

tion. The density of photoreceptors, bipolar cells, and

ganglion cells is highest at the fovea (Figure 22–1B).

The spacing between photoreceptors there is well

matched to the size of the optical blur circle, and thus

the image is sampled in an ideal fashion. Light must

generally traverse several layers of cells before reach-

ing the photoreceptors, but in the center of the fovea,

called the foveola, the other cellular layers are pushed

aside to reduce additional blur from light scattering

(Figure 22–1B). Finally, the back of the eye is lined by a

black pigment epithelium that absorbs light and keeps

it from scattering back into the eye.

The retina contains another special site, the optic disc,

where the axons of retinal ganglion cells converge and

extend through the retina to emerge from the back of

the eye as the optic nerve (Figure 22–1A). By neces-

sity, this area is devoid of photoreceptors and thus

Kandel-Ch22_0521-0544.indd 523 09/12/20 4:49 PM

图 22.0.2: 视网膜包括 5 个不同的神经元和突触层。A. 通过光学显微镜看到的人类视网膜的垂直截面。3 层细胞

体明显。外核层包含感光细胞体；内核层包括水平细胞、双极细胞和无长突细胞；神经节细胞层含有神经节细胞

和一些移位的无长突细胞。2 层纤维和突触将它们分开：外网层和内网层

[178]

。B. 基于高尔基体染色的猕猴视网

膜神经元。细胞层和突触层与 A 部分中的图像对齐

[179]

。

456

22.1 光感层对视觉图像进行采样

524 Part IV / Perception

Figure 22–3 The retinal circuitry.

A.The circuitry for cone signals, showing the split into ON

cell and OFF cell pathways (see Figure 22–10) as well as the

pathway for lateral inhibition in the outer layer. Red arrows

indicate sign-preserving connections through electrical or

glutamatergic synapses. Gray arrows represent sign-inverting

connections through GABAergic, glycinergic, or glutamatergic

synapses.

B.Rod signals feed into the cone circuitry through AII amacrine

cells, where the ON and OFF cell pathways diverge.

A 视锥细胞信号回路 B 视杆细胞信号回路

侧抑制

给光

光感受器

水平细胞

双极细胞

无长突细胞

神经节细胞

撤光给光撤光

无长突

细胞

杆状

双极细胞

视锥细胞

视杆细胞

corresponds to a blind spot in the visual field of each

eye. Because the disc lies nasal to the fovea of each eye,

light coming from a single point never falls on both

blind spots simultaneously, so that normally we are

unaware of them. We can experience the blind spot by

using only one eye (Figure 22–4). The blind spot dem-

onstrates what blind people experience—not black-

ness, but simply nothing. This explains why damage to

the peripheral retina often goes unnoticed. It is usually

through accidents, such as bumping into an unnoticed

object, or through clinical testing that a deficit of sight

is revealed.

The blind spot is a necessary consequence of the

inside-out design of the retina, which has puzzled and

amused biologists for generations. The purpose of this

organization may be to enable the tight apposition of

photoreceptors with the retinal pigment epithelium,

which plays an essential role in the turnover of retinal

pigment and recycles photoreceptor membranes by

phagocytosis.

There Are Two Types of Photoreceptors:

Rods and Cones

All photoreceptor cells have a common structure with

four functional regions: the outer segment, located at

the distal surface of the neural retina; the inner seg-

ment, located more proximally; the cell body; and the

synaptic terminal (Figure 22–5A).

Most vertebrates have two types of photorecep-

tors, rods and cones, distinguished by their morphol-

ogy. A rod has a long, cylindrical outer segment within

which the stacks of discs are separated from the plasma

membrane, whereas a cone often has a shorter, tapered

outer segment, and the discs are continuous with the

outer membrane (Figure 22–5B).

Kandel-Ch22_0521-0544.indd 524 09/12/20 4:49 PM

图 22.0.3: 视网膜回路。A. 视锥信号的回路，显示分裂成给光细胞通路和撤光细胞通路（参见图 22.3.1）以及外

层侧抑制通路。红色箭头表示通过电突触或谷氨酸能突触的信号保持连接。灰色箭头表示通过γ− 氨基丁酸能突

触、甘氨酸能突触或谷氨酸能突触的符号反转连接。B. 视杆信号通过 AII 无长突细胞进入视锥回路，给光细胞

通路和撤光细胞通路在此处分开。

457

22.1 光感层对视觉图像进行采样

22.1.1 眼科光学限制了视网膜图像的质量

视网膜图像的清晰度由几个因素决定：瞳孔孔径处的衍射、角膜和晶状体的屈光不正以及光路中材料引起

的散射。外界的一个点通常会聚焦在视网膜上一个模糊的小圆圈中。与其他光学设备一样，这种模糊在光轴附

近最小，图像质量接近光瞳衍射所施加的极限。远离轴，由于角膜和晶状体的像差，图像会显著退化，并且可能

会因光散射白内障或近视等屈光不正等异常情况而进一步退化。

靠近光轴的视网膜区域（中央凹）是视力最敏锐的地方，对应于指向我们关注物体的注视中心。如图 22.0.1B 所

示，感光细胞、双极细胞和神经节细胞的密度在中央凹处最高。感光器之间的间距与光学模糊圈的大小非常匹

配，因此以理想的方式对图像进行采样。如图 22.0.1B 所示，光在到达感光器之前通常必须穿过几层细胞，但在

称为中央凹的中央凹中心，其他细胞层被推到一边以减少光散射造成的额外模糊。最后，眼睛后部衬有一层黑

色色素上皮，它吸收光线并防止光线散射回眼睛。

如图 22.0.1A 所示，视网膜包含另一个特殊部位，即视盘，视网膜神经节细胞的轴突汇聚并延伸穿过视网

膜，从眼睛后部出现，成为视神经。必然地，该区域没有光受体，因此对应于每只眼睛视野中的盲点。由于圆盘

位于每只眼睛的中央凹的鼻侧，来自 1 个点的光不会同时落在 2 个盲点上，因此通常我们不会注意到它们。如

图 22.1.1 所示，我们可以只用一只眼睛体验盲点。盲点展示了盲人的体验（不是黑暗，而是什么都没有）。这就

解释了为什么周围视网膜的损伤经常被忽视。通常是由于意外事故，例如撞到未引起注意的物体，或通过临床

测试发现视力缺陷。

Chapter 22 / Low-Level Visual Processing: The Retina 525

Figure 22–4 The blind spot of the human retina.Locate the

blind spot in your left eye by shutting the right eye and fixating

the cross with the left eye. Hold the book about 12 inches from

your eye and move it slightly nearer or farther until the circle on

the left disappears. Now place a pencil vertically on the page

and sweep it sideways over the circle. Note the pencil appears

unbroken, even though no light can reach your retina from

the region of the circle. Next, move the pencil lengthwise and

observe what happens when its tip enters the circle. (Adapted,

with permission, from Hurvich 1981.)

Figure 22–5 Rod and cone photoreceptors have similar

structures.

A.Both rod and cone cells have specialized regions called the

outer and inner segments. The outer segment is attached to

the inner segment by a cilium and contains the light-transducing

apparatus. The inner segment holds mitochondria and much of

the machinery for protein synthesis.

B.The outer segment consists of a stack of membranous discs

that contain the light-absorbing photopigments. In both types of

cells, these discs are formed by infolding of the plasma mem-

brane. In rods, however, the folds pinch off from the membrane

so that the discs are free-floating within the outer segment,

whereas in cones, the discs remain part of the plasma membrane.

AAAS; Young 1970.)

A Morphology of photoreceptors B Outer segment of photoreceptors

Outer

segment

Inner

segment

Axon and

synaptic

terminal

Cell body

Cilium

Mitochondria

Nucleus

Discs

ConeRod

Free-ﬂoating

discs

Cytoplasmic

space

ConeRod

Folding of

outer cell

membrane

Connecting

cilium

Folding of

outer cell

membrane

Rods and cones also differ in function, most impor-

tantly in their sensitivity to light. Rods can signal the

absorption of a single photon and are responsible for

vision under dim illumination such as moonlight. But

as the light level increases toward dawn, the electri-

cal response of rods becomes saturated and the cells

cease to respond to variations in intensity. Cones are

much less sensitive to light; they make no contribution

to night vision but are solely responsible for vision in

daylight. Their response is considerably faster than

that of rods. Primates have only one type of rod but

three kinds of cone photoreceptors, distinguished by

the range of wavelengths to which they respond: the

L (long-wave), M (medium-wave), and S (short-wave)

cones (Figure 22–6).

The human retina contains approximately 100 million

rods and 5 million cones, but the two cell types are

differently distributed. The central fovea contains no

rods but is densely packed with small cones. A few

millimeters outside the fovea, rods greatly outnum-

ber cones. All photoreceptors become larger and more

widely spaced toward the periphery of the retina.

Kandel-Ch22_0521-0544.indd 525 09/12/20 4:49 PM

图 22.1.1: 人类视网膜的盲点。通过闭上右眼并用左眼固定十字来定位左眼的盲点。将书拿在离眼睛大约 12 英

寸的地方，然后稍微靠近或靠近一点，直到左边的圆圈消失。现在将一支铅笔垂直放在页面上，然后将它从侧面

扫过圆圈。请注意，尽管没有光线可以从圆圈区域到达您的视网膜，但铅笔似乎没有折断。接下来，纵向移动铅

笔，观察当它的笔尖进入圆圈时会发生什么

[180]

。

盲点是视网膜由内而外设计的必然结果，这让几代生物学家感到困惑和好笑。该组织的目的可能是使光受

体与视网膜色素上皮细胞紧密并置，这在视网膜色素更新和通过吞噬作用回收光受体膜中起着重要作用。

22.1.2 有 2 种类型的光受体：杆状和锥状

所有的感光细胞都有一个共同的结构，有四个功能区域：外节（位于神经视网膜的远端表面）、如图 22.1.2A 所

示，内段（位于更近端）、细胞体和突触末稍。

大多数脊椎动物有 2 种类型的光受体，即视杆细胞和视锥细胞，以它们的形态来区分。如图 22.1.2B 所示，

杆有一个长的圆柱形外段，其中圆盘堆叠与质膜分开，而圆锥通常有一个较短的锥形外段，圆盘与外膜连续。

视杆细胞和视锥细胞在功能上也有所不同，最重要的是它们对光的敏感性。视杆细胞可以发出吸收单个光

子的信号，并负责在月光等昏暗光照下的视觉。但是随着接近黎明的光照水平增加，杆状细胞的电反应变得饱

和，细胞停止对强度变化做出反应。视锥细胞对光的敏感度要低得多；它们对夜视没有贡献，但只负责白天的视

力。它们的反应比杆状体快得多。如图 22.2.3 所示，灵长类动物只有一种杆状体，但有 3 种锥形光受体，根据它

们响应的波长范围来区分：长波、中波和短波锥体。

人类视网膜包含大约 1 亿个视杆细胞和 500 万个视锥细胞，但这 2 种细胞类型的分布不同。中央凹不包含

杆状体，但密集地排列着小锥体。在中央凹外几毫米处，视杆细胞的数量远远多于视锥细胞。所有的光受体都变

得更大，并向视网膜周边分布得更广。短波锥体仅占所有锥体的 10%，并且不存在于中央凹中。

458

22.2 光转导将光子的吸收与膜电导的变化联系起来

Chapter 22 / Low-Level Visual Processing: The Retina 525

Figure 22–4 The blind spot of the human retina.Locate the

blind spot in your left eye by shutting the right eye and fixating

the cross with the left eye. Hold the book about 12 inches from

your eye and move it slightly nearer or farther until the circle on

the left disappears. Now place a pencil vertically on the page

and sweep it sideways over the circle. Note the pencil appears

unbroken, even though no light can reach your retina from

the region of the circle. Next, move the pencil lengthwise and

observe what happens when its tip enters the circle. (Adapted,

with permission, from Hurvich 1981.)

Figure 22–5 Rod and cone photoreceptors have similar

structures.

A.Both rod and cone cells have specialized regions called the

outer and inner segments. The outer segment is attached to

the inner segment by a cilium and contains the light-transducing

apparatus. The inner segment holds mitochondria and much of

the machinery for protein synthesis.

B.The outer segment consists of a stack of membranous discs

that contain the light-absorbing photopigments. In both types of

cells, these discs are formed by infolding of the plasma mem-

brane. In rods, however, the folds pinch off from the membrane

so that the discs are free-floating within the outer segment,

whereas in cones, the discs remain part of the plasma membrane.

AAAS; Young 1970.)

A 光感受器形态学 B 光感受器外段

外段

内段

轴突和

突触终末

胞体

纤毛

线粒体

细胞核

膜盘

视锥细胞视杆细胞

自由漂浮的

膜盘

细胞质

空间

视锥细胞视杆细胞

细胞外膜

折叠

连接

纤毛

细胞外膜

折叠

Rods and cones also differ in function, most impor-

tantly in their sensitivity to light. Rods can signal the

absorption of a single photon and are responsible for

vision under dim illumination such as moonlight. But

as the light level increases toward dawn, the electri-

cal response of rods becomes saturated and the cells

cease to respond to variations in intensity. Cones are

much less sensitive to light; they make no contribution

to night vision but are solely responsible for vision in

daylight. Their response is considerably faster than

that of rods. Primates have only one type of rod but

three kinds of cone photoreceptors, distinguished by

the range of wavelengths to which they respond: the

L (long-wave), M (medium-wave), and S (short-wave)

cones (Figure 22–6).

The human retina contains approximately 100 million

rods and 5 million cones, but the two cell types are

differently distributed. The central fovea contains no

rods but is densely packed with small cones. A few

millimeters outside the fovea, rods greatly outnum-

ber cones. All photoreceptors become larger and more

widely spaced toward the periphery of the retina.

Kandel-Ch22_0521-0544.indd 525 09/12/20 4:49 PM

图 22.1.2: 视杆光受体和视锥光感光器具有相似的结构。A. 视杆细胞和视锥细胞都有称为外段和内段的特殊区

域。外段通过纤毛连接到内段并包含光转换装置。内部部分包含线粒体和许多蛋白质合成机器。B. 外段由一堆

含有吸光感光色素的膜盘组成。在这 2 种类型的细胞中，这些圆盘都是通过细胞质膜的折叠形成的。然而，在杆

状细胞中，褶皱从膜上夹断，因此圆盘在外段内自由漂浮，而在锥体中，圆盘仍然是细胞质膜的一部分

[181-182]

。

视网膜注视中心显然专用于白天视觉。中央凹中锥形光受体的密集堆积设定了我们视力的极限。事实上，如

图 22.0.1C 所示，我们可以在医生的视力表上读到的最小字母具有笔画，其图像在视网膜上只有 1 到 2 个锥体直

径宽，视角约为 1 分弧。到了晚上，由于没有视杆细胞，中央凹是盲的。天文学家知道，必须只看一颗昏暗恒星

的侧面才能看到它。夜间在森林里散步时，我们这些非天文学家往往会跟随白天的反射，直视可疑声音的来源。

神秘的是，这个物体消失了，只是在我们移开视线时跳回了我们的周边视野。

22.2 光转导将光子的吸收与膜电导的变化联系起来

与许多其他神经元一样，光受体的膜电位受膜电导与 Na

和 K

平衡的调节，其跨膜梯度由代谢活性泵维持

（第 9 章）。在黑暗中，Na

通过非选择性阳离子通道流入光受体，该通道由第二信使环鸟苷-3,5-单磷酸盐激活。

色素蛋白对光子的吸收启动了生化级联反应，最终降低了环鸟苷-3,5-单磷酸盐的浓度，从而关闭了环鸟苷-

3,5-单磷酸盐门控通道，使细胞更接近 K

平衡电位。如图 22.2.1 所示，通过这种方式，光使感光器超极化。在

这里，我们详细描述了这一系列事件。这些知识大部分来自对杆状细胞的研究，但锥体中的机制非常相似。

22.2.1 光激活光受体中的色素分子

视紫红质是视杆细胞中的视觉色素，有 2 种成分。蛋白质部分视蛋白嵌入椎间盘膜中，本身不吸收可见光。

如图 22.2.2A 所示，吸光部分，视黄醛，是一种小分子，其 11-顺式异构体与视蛋白的赖氨酸残基共价连接。视

网膜对光子的吸收导致它从 11-cis 翻转到全反式构型。这种反应是视觉中唯一依赖光的步骤。

视网膜分子形状的变化导致视蛋白的构象变化为激活状态，称为变视紫红质 II，从而触发光转导的第二步。

变视紫红质 II 不稳定，会在几分钟内分裂，产生视蛋白和游离的全反式视网膜。然后全反式视黄醛从视杆细胞

运送到色素上皮细胞，在那里它被还原为全反式视黄醇（维生素 A），即 11-顺式视黄醛的前体，随后被运回视

杆细胞。

因此，全反式视网膜是视觉系统中的重要化合物。它的前体，如维生素 A，不能由人类合成，因此必须作为

日常饮食的一部分。维生素 A 缺乏会导致夜盲症，如果不加以治疗，还会导致受体外节的退化并最终导致失明。

459

22.2 光转导将光子的吸收与膜电导的变化联系起来

Chapter 22 / Low-Level Visual Processing: The Retina 527

暗亮

谷氨酸释放

细胞去极化

视紫红质

失活

通道打开

谷氨酸释放减少

细胞超极化

视紫红质

激活

通道关闭

视紫

红质

磷酸

二酯酶

环鸟苷-3,5-单磷酸盐

门控 Na

通道

环鸟苷-3,5-

单磷酸盐

磷酸

二酯酶

环鸟苷-3,5-

单磷酸盐

暗

亮

鸟苷-3,5-

单磷酸盐

βγ

三磷酸

鸟苷

鸟苷酸环化酶

视紫

红质

转导蛋白

鸟苷酸环化酶

三磷酸

鸟苷

时间（秒）

00.2

恢复暗亮

L 锥

光电压（毫伏）光电压（毫伏）

C 对光的电压响应

A 光转导与神经信号传导

增加活动或注意力

减少活动或注意力

磷酸

二酯酶

视紫

红质

环鸟苷-3,5-单磷酸盐

鸟苷酸环化酶

βγ

杆

转导蛋白

–20

–8

光转导中的分子过程

光转导中的反应网络

Kandel-Ch22_0521-0544.indd 527 09/12/20 4:49 PM

图 22.2.1: 光转导。A. 视杆细胞对光有反应。外段视盘中的视紫红质分子吸收光子，导致细胞质膜中环鸟苷-3,5-单

磷酸盐门控通道关闭。这种通道关闭使膜超极化并降低神经递质谷氨酸的释放速率

[7]

。B. 1. 光转导中的分子过

程。环鸟苷-3,5-单磷酸盐由鸟苷酸环化酶从三磷酸鸟苷产生并被磷酸二酯酶水解。在黑暗中，磷酸二酯酶活性

低，环鸟苷-3,5-单磷酸盐浓度高，环鸟苷-3,5-单磷酸盐门控通道开放，允许 Na

和 Ca

流入。在光线下，视紫红

质通过吸收光子而被激发，然后激活转导蛋白，转导蛋白进而激活磷酸二酯酶；环鸟苷-3,5-单磷酸盐水平下降，

膜通道关闭，更少的 Na

和 Ca

进入细胞。转导酶均位于内膜圆盘中，可溶性配体环鸟苷-3,5-单磷酸盐作为质

膜的信使。2. Ca

在光转导反应级联中具有负反馈作用。光对网络的刺激导致环鸟苷-3,5-单磷酸盐门控通道的

关闭。这导致细胞内 Ca

浓度下降。因为 Ca

调节至少 3 个级联组分的功能（视紫红质、鸟苷酸环化酶和环鸟

苷-3,5-单磷酸盐门控通道）Ca

的下降抵消了光引起的激发。C. 灵长类视杆细胞和视锥细胞对强度增加的短暂

闪光的电压响应。迹线上的数字越高表示光照强度越大（并非所有迹线都已标记）。对于昏暗的闪光，响应幅度

随强度线性增加。在高强度下，受体饱和并在闪光后的一段时间内稳定地保持超极化；这会导致我们在明亮的

闪光后看到的残像。请注意，对于更亮的闪光，响应峰值更早，并且视锥细胞的响应速度比视杆细胞快

[183]

。

460

22.2 光转导将光子的吸收与膜电导的变化联系起来

528 Part IV / Perception

Figure 22–8 Structure of the visual pigments.

A.The visual pigment in rod cells, rhodopsin, is the covalent

complex of two components. Opsin is a large protein with

348 amino acids and a molecular mass of approximately

40,000 daltons. It loops back and forth seven times across

the membrane of the rod disc. Retinal is a small light-

absorbing compound covalently attached to a side chain of

lysine 296 in opsin’s seventh membrane-spanning region.

Absorption of light by 11-cis retinal causes a rotation around

the double bond. As retinal adopts the more stable all-trans

configuration, it causes a conformational change in opsin

that triggers the subsequent events of visual transduction.

(Adapted, with permission, from Nathans and Hogness

1984.)

B.The blue circles denote identical amino acids; black circles

denote differences. The forms of opsin in the three types of cone

cells (L, M, and S) resemble each other as well as the rhodopsin

in rod cells, suggesting that all four evolved from a common

precursor by duplication and divergence. The L and M opsins

are most closely related, with 96% identity in their amino acid

sequences. They are thought to have evolved from a gene-

duplication event approximately 30 million years ago, after Old

World monkeys, which have three visual pigments, separated

from New World monkeys, which generally have only two.

Light Activates Pigment Molecules in the

Photoreceptors

Rhodopsin, the visual pigment in rod cells, has two

components. The protein portion, opsin, is embedded

in the disc membrane and does not by itself absorb vis-

ible light. The light-absorbing moiety, retinal, is a small

molecule whose 11-cis isomer is covalently linked to a

细胞质

外段

A 视杆中的视觉色素

圆盘膜

Residue 296

（视网膜

附着部位）

COOH

视觉色素

（视紫红质）

圆盘内部

11-cis retinal (M

= 268)

全反式视黄醛

光

CCH

B 视觉色素氨基酸序列

M vs

视紫红质

视蛋白视网膜

COOH

M vs S

L vs M

COOH

细胞质

椎间盘

间隙

圆盘膜

圆盘内部

细胞质

圆盘膜

lysine residue of opsin (Figure 22–8A). Absorption of a

photon by retinal causes it to flip from the 11-cis to the

all-trans configuration. This reaction is the only light-

dependent step in vision.

The change in shape of the retinal molecule causes

a conformational change in the opsin to an activated

state called metarhodopsin II, thus triggering the second

Kandel-Ch22_0521-0544.indd 528 09/12/20 4:49 PM

图 22.2.2: 视觉色素的结构。A. 视杆细胞中的视色素视紫红质是 2 种成分的共价复合物。视蛋白是一种大型蛋白

质，具有 348 个氨基酸，分子量约为 4 万道尔顿。它在杆盘的膜上来回循环 7 次。视黄醛是一种小的吸光化合物，

共价连接到视蛋白第 7 个跨膜区域赖氨酸 296 的侧链上。11-顺式视黄醛对光的吸收导致围绕双键的旋转。由于

视网膜采用更稳定的全反式构型，它会引起视蛋白的构象变化，从而触发后续的视觉转导事件

[184]

。B. 蓝色圆圈

表示相同的氨基酸；黑色圆圈表示差异。3 种视锥细胞（长波、中波和短波中的视蛋白形式彼此相似，视杆细胞

中的视紫红质也相似，这表明这 4 种细胞都是通过复制和发散从一个共同的前体进化而来的。长波和中波视蛋

白的关系最为密切，其氨基酸序列具有 96% 的同一性。它们被认为是从大约 3 千万年前的一次基因复制事件中

进化而来的，当时旧大陆猴具有 3 种视觉色素，而新大陆猴通常只有 2 种。

461

22.2 光转导将光子的吸收与膜电导的变化联系起来

人类视网膜中的每种视锥细胞都会产生一种视蛋白变体。如图 22.2.3 所示，这 3 种锥形颜料的区别在于它

们的吸收光谱，即光吸收效率对波长的依赖性。光谱由蛋白质序列通过结合口袋附近的视黄醛和某些氨基酸侧

链之间的相互作用来确定。红光比中波视锥细胞更能激发长波视锥细胞，而绿光更能激发中波视锥细胞。因此，

这些锥体类型中的相对激发度包含有关光的光谱的信息，与其强度无关。大脑对来自不同视锥细胞类型的信号

进行比较是色觉的基础。

526 Part IV / Perception

Figure 22–6 Sensitivity spectra for the three types of cones

and the rod.At each wavelength, the sensitivity is inversely

proportional to the intensity of light required to elicit a criterion

response in the sensory neuron. Sensitivity varies over a large

range and thus is shown on a logarithmic scale. The different

classes of photoreceptors are sensitive to broad and overlap-

ping ranges of wavelengths. (Reproduced, with permission,

from Schnapf et al. 1988.)

Figure 22–7 (Opposite) Phototransduction.

A.The rod cell responds to light. Rhodopsin molecules in the

outer-segment discs absorb photons, which leads to the clo-

sure of cyclic guanosine 3′-5′ monophosphate (cGMP)-gated

channels in the plasma membrane. This channel closure hyper-

polarizes the membrane and reduces the rate of release of the

neurotransmitter glutamate. (Adapted from Alberts 2008.)

B.1. Molecular processes in phototransduction. cGMP is

produced by a guanylate cyclase (GC) from guanosine triphos-

phate (GTP) and hydrolyzed by a phosphodiesterase (PDE).

In the dark, the phosphodiesterase activity is low, the cGMP

concentration is high, and the cGMP-gated channels are open,

allowing the influx of Na

and Ca

. In the light, rhodopsin (R)

is excited by absorption of a photon, then activates transducin

(T), which in turn activates the PDE; the cGMP level drops, the

membrane channels close, and less Na

and Ca

enter the cell.

The transduction enzymes are all located in the internal mem-

brane discs, and the soluble ligand cGMP serves as a messen-

ger to the plasma membrane.

2. Calcium ions have a negative feedback role in the reaction

cascade in phototransduction. Stimulation of the network

by light leads to the closure of the cGMP-gated channels.

This causes a drop in the intracellular concentration of Ca

Because Ca

modulates the function of at least three com-

ponents of the cascade—rhodopsin, GC, and the cGMP-gated

channel—the drop in Ca

counteracts the excitation caused by

light.

C.Voltage response of a primate rod and cone to brief flashes

of light of increasing intensity. Higher numbers on the traces

indicate greater intensities of illumination (not all traces are

labeled). For dim flashes, the response amplitude increases

linearly with intensity. At high intensities, the receptor saturates

and remains hyperpolarized steadily for some time after the

flash; this leads to the afterimages that we perceive after a

bright flash. Note that the response peaks earlier for brighter

flashes and that cones respond faster than rods. (Reproduced,

with permission, from Schneeweis and Schnapf 1995. Copy-

–1

–2

–3

–4

–5

–6

400 500

S 视锥细胞

M 视锥

细胞

视锥细胞

视杆细胞

600

波长（纳米）

相对灵敏度的对数

700 800 900

The S cones make up only 10% of all cones and are

absent from the central fovea.

The retinal center of gaze is clearly specialized for

daytime vision. The dense packing of cone photorecep-

tors in the fovea sets the limits of our visual acuity. In

fact, the smallest letters we can read on a doctor’s eye

chart have strokes whose images are just one to two

cone diameters wide on the retina, a visual angle of

about 1 minute of arc (Figure 22–1C). At night, the cen-

tral fovea is blind owing to the absence of rods. Astron-

omers know that one must look just to the side of a dim

star to see it at all. During nighttime walks in the forest,

we nonastronomers tend to follow our daytime reflex

of looking straight at the source of a suspicious sound.

Mysteriously, the object disappears, only to jump back

into our peripheral field of view as we avert our gaze.

Phototransduction Links the Absorption of a

Photon to a Change in Membrane Conductance

As in many other neurons, the membrane potential of

a photoreceptor is regulated by the balance of mem-

brane conductances to Na

and K

ions, whose trans-

membrane gradients are maintained by metabolically

active pumps (Chapter 9). In the dark, Na

ions flow

into the photoreceptor through nonselective cation

channels that are activated by the second messenger

cyclic guanosine 3′-5′ monophosphate (cGMP).

Absorption of a photon by the pigment protein

sets in motion a biochemical cascade that ultimately

lowers the concentration of cGMP, thus closing the

cGMP-gated channels and moving the cell closer to the

equilibrium potential. In this way, light hyperpolar-

izes the photoreceptor (Figure 22–7). Here, we describe

this sequence of events in detail. Most of this knowl-

edge derives from studies of rods, but the mechanism

in cones is very similar.

Kandel-Ch22_0521-0544.indd 526 09/12/20 4:49 PM

图 22.2.3: 3 种锥体和杆体的灵敏度光谱。在每个波长下，灵敏度与在感觉神经元中引起标准反应所需的光强度

成反比。灵敏度在很大范围内变化，因此以对数刻度显示。不同类别的光受体对广泛且重叠的波长范围敏感。

在夜视中，只有视杆细胞处于活动状态，因此所有功能性光受体都具有相同的吸收光谱。因此，绿光对视觉

系统的影响与强度更高的红光完全相同。因为单光受体系统无法区分光的光谱和强度，所以“晚上所有的猫都

是灰色的”。通过比较杆对不同波长光的敏感度，可以获得视紫红质的吸收光谱。如图 22.2.4 所示，一个了不起

的事实是，只需向人类受试者询问各种彩色光的外观，就可以准确地测量这种分子特性。知觉或心理物理学的

定量研究为大脑处理的其他机制提供了类似的见解（第 17 章）。

22.2.2 兴奋的视紫红质通过 G 蛋白转导蛋白激活磷酸二酯酶

活化的视紫红质以变视紫红质 II 的形式扩散到视盘膜内，在那里它遇到转导蛋白，G 蛋白家族的一员（第 14 章）。

与其他 G 蛋白的情况一样，无活性形式的转导蛋白结合二磷酸鸟苷分子。与变视紫红质 II 的相互作用促进二磷

酸鸟苷与三磷酸鸟苷的交换。这导致转导蛋白的亚基解离为携带三磷酸鸟苷（Tα-三磷酸鸟苷）以及 β 和 γ 亚基

（Tβγ）的活性 α 亚基。变视紫红质 II 可以激活数百个额外的转导蛋白分子，从而显著增强细胞的反应。

活性转导蛋白亚基 Tα-三磷酸鸟苷与环核苷酸磷酸二酯酶（另一种与椎间盘膜相关的蛋白质）形成复合物。

这种相互作用大大提高了酶将环鸟苷-3,5-单磷酸盐水解为 5’-鸟苷-3,5-单磷酸盐的速率。每个磷酸二酯酶分子每

秒可水解 1 千多个环鸟苷-3,5-单磷酸盐分子，从而增加放大程度。

环鸟苷-3,5-单磷酸盐的浓度控制外段质膜中环鸟苷-3,5-单磷酸盐门控通道的活性。在黑暗中，当环鸟苷-3,5-

单磷酸盐浓度高时，大量 Na

通过开放通道流入，使细胞保持在大约–40 毫伏的去极化水平。结果，细胞的突触

末稍不断释放递质谷氨酸。如图 22.2.1B1 所示，光诱发的环鸟苷-3,5-单磷酸盐减少导致环鸟苷-3,5-单磷酸盐门

控通道关闭，从而减少 Na

的内流并使细胞超极化。超极化减缓了光受体末端神经递质的释放，从而启动了神

经信号。

462

22.2 光转导将光子的吸收与膜电导的变化联系起来

Chapter 22 / Low-Level Visual Processing: The Retina 529

Figure 22–9 Absorption spectrum of rhodopsin. The

absorption spectrum of human rhodopsin measured in a

cuvette is compared with the spectral sensitivity of human

observers to very dim light flashes. The psychophysical

data have been corrected for absorption by the ocular

media. (Reproduced, with permission, from Wald and

step of phototransduction. Metarhodopsin II is unsta-

ble and splits within minutes, yielding opsin and free

all-trans retinal. The all-trans retinal is then trans-

ported from rods to pigment epithelial cells, where it

is reduced to all-trans retinol (vitamin A), the precur-

sor of 11-cis retinal, which is subsequently transported

back to rods.

All-trans retinal is thus a crucial compound in the

visual system. Its precursors, such as vitamin A, can-

not be synthesized by humans and so must be a regu-

lar part of the diet. Deficiencies of vitamin A can lead

to night blindness and, if untreated, to deterioration of

receptor outer segments and eventually to blindness.

Each type of cone in the human retina produces a var-

iant of the opsin protein. These three cone pigments are

distinguished by their absorption spectrum, the depend-

ence on wavelength of the efficiency of light absorption

(see Figure 22–6). The spectrum is determined by the

protein sequence through the interaction between reti-

nal and certain amino acid side chains near the binding

pocket. Red light excites L cones more than the M cones,

whereas green light excites the M cones more. Therefore,

the relative degree of excitation in these cone types con-

tains information about the spectrum of the light, inde-

pendent of its intensity. The brain’s comparison of signals

from different cone types is the basis for color vision.

In night vision, only the rods are active, so all

functional photoreceptors have the same absorption

spectrum. A green light consequently has exactly the

same effect on the visual system as a red light of a

greater intensity. Because a single-photoreceptor sys-

tem cannot distinguish the spectrum of a light from its

intensity, “at night all cats are gray.” By comparing the

sensitivity of a rod to different wavelengths of light,

one obtains the absorption spectrum of rhodopsin. It is

视紫红质吸收光谱

人类暗亮光谱灵敏度

波长（纳米）

1.0

0.8

0.6

0.4

0.2

0.0

400 450 500 550 600650 70

灵敏度

a remarkable fact that one can measure this molecu-

lar property accurately just by asking human sub-

jects about the appearance of various colored lights

(Figure 22–9). The quantitative study of perception,

or psychophysics, provides similar insights into other

mechanisms of brain processing (Chapter 17).

Excited Rhodopsin Activates a Phosphodiesterase

Through the G Protein Transducin

Activated rhodopsin in the form of metarhodopsin II

diffuses within the disc membrane where it encounters

transducin, a member of the G protein family (Chapter 14).

As is the case for other G proteins, the inactive form of

transducin binds a molecule of guanosine diphosphate

(GDP). Interaction with metarhodopsin II promotes the

exchange of GDP for guanosine triphosphate (GTP).

This leads to dissociation of transducin’s subunits into

an active α-subunit carrying the GTP (Tα-GTP) and the

β- and γ-subunits (Tβγ). Metarhodopsin II can activate

hundreds of additional transducin molecules, thus sig-

nificantly amplifying the cell’s response.

The active transducin subunit Tα-GTP forms a

complex with a cyclic nucleotide phosphodiesterase,

another protein associated with the disc membrane.

This interaction greatly increases the rate at which the

enzyme hydrolyzes cGMP to 5′-GMP. Each phosphodi-

esterase molecule can hydrolyze more than 1,000 mol-

ecules of cGMP per second, thus increasing the degree

of amplification.

The concentration of cGMP controls the activity of

the cGMP-gated channels in the plasma membrane of

the outer segment. In darkness, when the cGMP con-

centration is high, a sizeable Na

influx through the

open channels maintains the cell at a depolarized level

Kandel-Ch22_0521-0544.indd 529 09/12/20 4:49 PM

图 22.2.4: 视紫红质的吸收光谱。将在比色皿中测量的人类视紫红质的吸收光谱与人类观察者对非常暗淡的闪光

的光谱灵敏度进行比较。心理物理数据已针对眼部介质的吸收进行了校正。

22.2.3 多重机制关断级联

感光器对单个光子的反应必须终止，这样细胞才能对另一个光子做出反应。变视紫红质 II 通过特定视紫红

质激酶的磷酸化作用失活，然后结合可溶性蛋白抑制蛋白，从而阻断与转导蛋白的相互作用。

活性转导蛋白（Tα-三磷酸鸟苷）具有内在的三磷酸鸟苷酶活性，最终将结合的三磷酸鸟苷转化为二磷酸鸟

苷。然后 Tα-二磷酸鸟苷释放磷酸二酯酶并与 Tβγ 重组，再次准备好被视紫红质激发。一旦磷酸二酯酶失活，环

鸟苷-3,5-单磷酸盐浓度就会通过从三磷酸鸟苷产生环鸟苷-3,5-单磷酸盐的鸟苷酸环化酶恢复。此时，膜通道打

开，Na

电流恢复，光受体去极化回到其暗电位。

除了这些关闭级联的各个元素的独立机制之外，重要的反馈机制可确保更快地终止大量响应。这是由细胞

中 Ca

浓度的变化介导的。Ca

通过环鸟苷-3,5-单磷酸盐门控通道进入细胞，并被快速阳离子交换剂排出。在

黑暗中，细胞内 Ca

浓度较高，但在细胞对光反应期间，当环鸟苷-3,5-单磷酸盐门控通道关闭时，Ca

水平迅

速下降至黑暗水平的百分之几。

如图 22.2.1B2 所示，Ca

浓度的降低以 3 种方式调节生化反应。钙结合蛋白恢复素对视紫红质激酶的作用

可加速视紫红质磷酸化，从而减少转导蛋白的激活。钙依赖性鸟苷酸环化酶激活蛋白可加速鸟苷酸环化酶的活

性。最后，环鸟苷-3,5-单磷酸盐门控通道对环鸟苷-3,5-单磷酸盐的亲和力通过 Ca

-钙调蛋白的作用而增加。所

有这些作用促进光受体返回到黑暗状态。

22.2.4 光转导缺陷导致疾病

毫不奇怪，光转导机制中的缺陷会产生严重的后果。一个突出的缺陷是色盲，这是由于视锥细胞色素基因

缺失或异常引起的，稍后将对此进行讨论。

当视杆细胞功能丧失但视锥细胞功能保持完好时，就会导致静止性夜盲症。这种疾病是可遗传的，并且已

经在光转导级联的许多组分中发现了突变：视紫红质、杆状转导蛋白、杆状磷酸二酯酶、视紫红质激酶和抑制蛋

白。在某些情况下，视杆细胞似乎会永久激活，就好像暴露在持续的致盲光下一样。

不幸的是，许多光转导缺陷会导致色素性视网膜炎，这是一种视网膜进行性退化，最终导致失明。该疾病有

463

22.3 神经节细胞将神经图像传输到大脑

多种形式，其中许多与影响视杆细胞信号转导的突变有关。为什么这些功能变化会导致视杆细胞死亡以及随后

的视锥细胞退化尚不清楚。

22.3 神经节细胞将神经图像传输到大脑

感光层产生视觉场景的相对简单的神经表征：明亮区域的神经元超极化，而黑暗区域的神经元去极化。由

于视神经的轴突数量仅为受体细胞数量的 1%，因此视网膜回路必须在将信息传送到大脑之前编辑光受体中的信

息。

此步骤构成低层视觉处理，是从落在视网膜上的光模式中获得视觉感知的第一阶段。要了解这个过程，我

们必须首先了解视网膜输出的组织以及视网膜神经节细胞如何响应各种光模式。

22.3.1 神经节细胞的 2 种主要类型是给光细胞和撤光细胞

许多视网膜神经节细胞即使在黑暗或持续光照下也会自发地激发动作电位。如果光强度突然增加，所谓的

给光细胞会更快地放电。其他神经节细胞，撤光细胞，放电更慢或完全停止放电。当强度再次降低时，给光细胞

的放电减少而撤光细胞的放电增加。因此，视网膜输出包括 2 个互补的表示，它们对光的反应极性不同。

这种布置用于快速传达视觉场景中的变亮和变暗。如果视网膜只有给光细胞，黑暗物体将通过降低放电率

来编码。如果神经节细胞以每秒 10 个脉冲的维持速率放电，然后降低其速率，则突触后神经元需要大约 100 毫

秒才能注意到动作电位频率的变化。相比之下，激活率增加到每秒 200 个脉冲仅在 5 毫秒内就很明显。

22.3.2 许多神经节细胞对图像中的边缘反应强烈

为了更详细地探测神经节细胞的反应，可以测试细胞的放电如何随着聚焦在视网膜不同部分的小光点的位

置和时间过程而变化。

典型的神经节细胞对靠近细胞体的视网膜紧凑区域（称为细胞感受野）中的光敏感。在那个区域内，人们通

常可以区分中心区域和周围区域，光在细胞中产生相反的反应。例如，当亮点聚焦在细胞的感受野中心时，给光

细胞会激活得更快，但当亮点聚焦在周围时会减少发射。如果光线覆盖中心和周围，则响应比仅中心照明弱得

多。中心的亮点与覆盖周围的黑色环形相结合，会引发非常强烈的激活。如图 22.3.1 所示，对于撤光细胞，这些

关系是相反的；细胞被暗点和亮环强烈激发。

因此，由一群视网膜神经节细胞产生的输出增强了输入中的空间对比度区域，例如 2 个不同强度区域之间

的边缘，并且不太强调均匀照明区域。

22.3.3 神经节细胞的输出强调刺激的时间变化

当出现有效的光刺激时，神经节细胞的放电通常会从静息水平急剧增加到峰值，然后放松到中间速率。当

刺激关闭时，放电率急剧下降，然后逐渐恢复到静止水平。

从峰值到静息水平下降的速度因神经节细胞类型而异。如图 22.3.1 所示，瞬时神经元仅在刺激开始时产生

脉冲，而持续神经元在刺激期间保持几乎稳定的放电率几秒钟。

然而，一般来说，神经节细胞的输出有利于视觉输入在恒定光强度期间的时间变化。事实上，当图像通过眼动

跟踪设备稳定在视网膜上时，它会在几秒钟内从视野中消失。幸运的是，这在正常视力中永远不会发生；即使当我

们试图固定我们的视线时，小的自动眼球运动（眼跳）也会持续扫描视网膜上的图像并防止世界消失（第 25 章）。

22.3.4 视网膜输出强调移动物体

基于这些观察，我们可以更普遍地理解神经节细胞对视觉输入的反应。例如，如图 22.3.2 所示，移动物体的

边缘会在神经节细胞群中引起强烈的放电，因为这些是空间对比的唯一区域，也是光强度随时间变化的唯一区

域。

464

22.3 神经节细胞将神经图像传输到大脑

532 Part IV / Perception

刺激模式激活率

210

时间（秒）

210

时间（秒）

刺激

撤光细胞给光细胞

4 中心和周围

相反

3 中心和周围

一致

2 仅周围

1 仅中心

瞬态细胞维持的细胞

刺激

区域内

区域外

Figure 22–10 Responses of retinal ganglion cells with

center-surround receptive fields.In these idealized experi-

ments, the stimulus changes from a uniform gray field to the

pattern of bright (yellow) and dark (black) regions indicated

on the left. This leads to the firing rate responses shown

on the right. 1. ON cells are excited by a bright spot in the

receptive field center, OFF cells by a dark spot. In

sustained cells, the excitation persists throughout

stimulation, whereas in transient cells, a brief burst of spikes

occurs just after the onset of stimulation. 2. If the same

stimulus that excites the center is applied to the surround,

firing is suppressed. 3. Uniform stimulation of both center

and surround elicits a response like that of the center, but

much smaller in amplitude. 4. Stimulation of the center com-

bined with the opposite stimulus in the surround produces

the strongest response.

Kandel-Ch22_0521-0544.indd 532 09/12/20 4:49 PM

图 22.3.1: 具有中心环绕感受野的视网膜神经节细胞反应。在这些理想化的实验中，刺激从均匀的灰色区域变为

左侧指示的明亮（黄色）和黑暗（黑色）区域的模式。这导致右侧显示的激活率响应。1. 给光细胞被感受野中

心的亮点激发，撤光细胞被暗点激发。在持续细胞中，兴奋在整个刺激过程中持续存在，而在瞬态细胞中，在刺

激开始后会出现短暂的脉冲。2. 如果将激发中心的相同刺激施加到周围，则会抑制放电。3. 中心和周围的均匀

刺激引起与中心类似的反应，但幅度小得多。4. 中心的刺激与周围的相反刺激相结合会产生最强的反应。

465

22.3 神经节细胞将神经图像传输到大脑

Chapter 22 / Low-Level Visual Processing: The Retina 533

Figure 22–11 Responses of ganglion cells in the cat retina

to moving objects.

A.The firing rate of an ON ganglion cell in response to a vari-

ety of bars (white or black, various widths) moving across the

retina. Each bar moves at 10° per second; 1° corresponds to

180 μm on the retina. In response to the white bar, the firing

rate first decreases as the bar passes over the receptive-field

surround (1), increases as the bar enters the center (2), and

decreases again as the bar passes through the surround on

the opposite side (3). The dark bar elicits responses of the

opposite sign. Because ganglion cells similar to this one are

distributed throughout the retina, one can also interpret this

curve as an instantaneous snapshot of activity in a population

of ganglion cells, where the horizontal axis represents loca-

tion on the retina. In effect, this activity profile is the neural

representation of the moving bar transmitted to the brain. A

complementary population of OFF ganglion cells (not shown

here) conveys another neural activity profile in parallel. In this

way, both bright and dark edges can be signaled by a sharp

increase in firing.

B.A simple model of retinal processing that incorporates

center-surround antagonism and a transient temporal filter is

used to predict ganglion-cell firing rates. The predictions match

the essential features of the responses in part A. (Reproduced,

Elsevier Ltd.)

50 脉冲/秒

1 s

2°

5°

0.5°

白条黑条

A 给光细胞响应

B 模型预测

Kandel-Ch22_0521-0544.indd 533 09/12/20 4:49 PM

图 22.3.2: 猫视网膜神经节细胞对移动物体的反应。A. 给光神经节对各种横条（白色或黑色，不同宽度）在视网

膜上移动的响应率。每个条以每秒 10° 的速度移动；1° 对应于视网膜上的 180 微米。作为对白色条的响应，当

条穿过感受野环绕（1）时，激活率首先降低，随着条进入中心（2）而增加，当条穿过对面的环绕时再次降低

（3）。深色条引起相反符号的反应。由于与此类似的神经节细胞分布在整个视网膜中，因此也可以将此曲线解释

为神经节细胞群活动的瞬时快照，其中横轴表示在视网膜上的位置。实际上，此活动配置文件是传输到大脑的

移动杆的神经表征。一个互补的撤光神经节细胞群（此处未显示）并行传达另一个神经活动概况。这样，亮边和

暗边都可以通过激活的急剧增加来表示。B. 一个简单的视网膜处理模型，结合了中心环绕拮抗作用和瞬态时间

过滤器，用于预测神经节细胞放电率。这些预测与 A 部分中的响应的基本特征相匹配。

466

22.3 神经节细胞将神经图像传输到大脑

我们很容易理解为什么视网膜会选择性地响应这些特征。物体的轮廓对于推断其形状和身份特别有用。同

样，突然移动或改变的物体比那些不移动或改变的物体更值得立即关注。因此，视网膜处理提取了有助于指导

行为的场景低层特征，并有选择地将这些特征传输到大脑。事实上，拒绝在空间或时间上保持不变的特征解释

了人类感知的时空敏感性（文本框 22.1）。

文本框 22.1 (人类感知的时空敏感性)

♠

尽管小光斑有助于探测视觉通路中单个神经元的感受野，但了解人类视觉感知需要不同的刺激。光

栅刺激通常用于探测我们的视觉系统如何处理空间和时间模式。

如图 22.5.4 所示，受试者将强度在平均值附近变化的显示器视为空间的正弦函数。然后，显示器的对

比度（定义为正弦曲线的峰间振幅除以平均值）降低到光栅几乎看不见的阈值。对于不同空间频率的光

栅重复该测量。

如图 22.5.3A 所示，当根据空间频率绘制该阈值的倒数时，由此产生的对比敏感度曲线提供了视觉感

知对不同尺度模式的敏感度的测量。当在高光强度下测量时，灵敏度在高空间频率下急剧下降，绝对阈值

约为每度 50 个周期。如图 22.0.1C 所示，这种灵敏度从根本上受到光学图像质量和中央凹锥细胞间距的

限制。

有趣的是，在较低的空间频率下，灵敏度也会下降。频率约为每度 5 个周期的图案最为明显。视觉系

统被认为具有带通行为，因为它除了一个空间频率带之外，拒绝所有的空间频率。

人们可以使用同样的技术来测量灵长类动物单个视网膜神经节细胞的敏感性。如图 22.5.3 所示，结

果与人类受试者的结果相似，表明视觉感知的这些基本特征是由视网膜决定的。

带通行为可以基于中心-周围感受野的空间对抗来理解。一个非常精细的光栅在感受野中心呈现出许

多明暗条纹；它们的作用相互抵消，因此不提供净激励。一个非常粗糙的光栅在感受野的中心和周围都

呈现出单一的条纹，它们的拮抗作用再次为神经节细胞提供了少量的净兴奋。如图 22.5.3B 所示，最强的

响应是由中等空间频率的光栅产生的，该光栅仅用一条条纹覆盖中心，并用相反极性的条纹覆盖大部分

周围。

如图 22.5.3A 所示，在昏暗的光线下，视觉系统的对比敏感度下降，但在高空间频率下比在低空间频

率下更为严重。因此，峰值灵敏度转移到较低的空间频率，并且最终曲线完全失去其峰值。在这种状态

下，视觉系统具有所谓的低通行为，因为它优先编码低空间频率的刺激。如图 22.5.3B 所示，在昏暗的光

线下，神经节细胞的感受野失去了它们的拮抗环境，这一事实解释了从带通到低通空间滤波的转变。

可以进行类似的实验来测试视觉对时间模式的敏感性。这里，测试刺激的强度随时间呈正弦曲线闪

烁，而对比度逐渐达到检测的阈值水平。如图 22.4.2A

所示，对于人类来说，对比敏感度在非常高的闪烁

频率下急剧下降，但在非常低的频率下也会下降。大约 10 赫兹的闪烁是最有效的刺激。如图 22.4.2B 所

示，人们在猕猴视网膜神经节细胞的闪烁敏感性中发现了类似的带通行为。

对时间对比度的敏感度也取决于平均光照水平。如图 22.4.2 所示，对于人类受试者，最佳闪烁频率在

较低的刺激强度下向下移动，曲线中的峰值变得越来越不突出。灵长类动物视网膜神经节细胞重复这种

行为的事实表明，视网膜处理限制了整个视觉系统在这些简单任务中的表现。

22.3.5 几种神经节细胞类型通过并行通路投射到大脑

根据它们的形态和对光的反应，已经鉴定出几种不同类型的神经节细胞。如图 22.0.2B 所示，给光细胞和撤

光细胞出现在每个脊椎动物的视网膜中，而在灵长类动物的视网膜中，两大类细胞，小细胞和大细胞，分别包

括给光和撤光类型。在距中央凹的任何给定距离处，大细胞的感受野远大于小细胞的感受野。大细胞也比小细

胞具有更快和更瞬态的反应。由于视觉色素黑视蛋白的表达，一些神经节细胞本质上对光敏感。

总共描述了 20 多种类型的神经节细胞。每种类型的细胞群以平铺方式覆盖视网膜，使得视网膜上的任何点

位于至少一个神经节细胞的感受野中心内。可以设想来自每个群体的信号一起向大脑发送视野的独特神经表征。

在这个观点中，视神经传达了 20 种或更多的神经表征，它们在极性（开或关）、空间分辨率（精细或粗糙）、时

467

22.4 中间神经元网络塑造视网膜输出

间响应（持续或瞬态）、光谱过滤（宽带或由红色、绿色或蓝色），以及对其他图像特征（例如运动）的选择性。

这些神经表征指向大脑中的各种视觉中心，包括丘脑的外侧膝状体核，它是视觉皮层的中继；上丘，一个

涉及空间注意力和定向运动的中脑区域；前顶盖参与对瞳孔的控制；辅助光学系统，分析自身运动以稳定凝视；

和视交叉上核，一个指导昼夜节律的中央时钟，其相位可以通过光提示设置（第 44 章）。在许多情况下，一种神

经节细胞的轴突将侧枝延伸到中枢神经系统的多个区域。例如，大细胞投射到丘脑和上丘。

22.4 中间神经元网络塑造视网膜输出

我们现在更详细地考虑视网膜回路以及它如何解释视网膜神经节细胞的复杂反应特性。

22.4.1 并行通路起源于双极细胞

如图 22.0.3A 所示，光受体与双极细胞和水平细胞形成突触。在黑暗中，光受体的突触末稍不断释放谷氨酸。

当受光刺激时，光受体超极化，较少的钙进入末端，末端释放较少的谷氨酸。光受体不发射动作电位；像双极细

胞一样，它们使用特殊结构（带状突触）以分级方式释放神经递质。事实上，大多数视网膜处理是通过分级膜电

位完成的：动作电位主要发生在某些无长突细胞和视网膜神经节细胞中。

双极细胞的 2 种主要变体，给光细胞和撤光细胞，通过不同的机制对突触处的谷氨酸作出反应。撤光细胞

使用离子型受体，即α-氨基-3-羟基-5-甲基-4-异恶唑丙酸-红藻氨酸变体的谷氨酸门控阳离子通道。在黑暗中释放

的谷氨酸使这些细胞去极化。给光细胞使用与 G 蛋白相连的促代谢受体，G 蛋白的作用最终会关闭阳离子通道。

因此，这些受体的谷氨酸激活使黑暗中的细胞超极化。

Chapter 22 / Low-Level Visual Processing: The Retina 537

Figure 22–15 Bipolar cells in the macaque retina. The cells

are arranged according to the depth of their terminal arbors in

the inner plexiform layer. The horizontal line dividing the distal

(upper) and proximal (lower) levels of this layer represents

the border between the axonal terminals of OFF and ON type

cells. Terminals in the upper half are presumed to be those of

OFF cells, and those in the lower half ON cells. Cell types are

diffuse bipolar cells (DB), ON and OFF midget bipolars (IMB,

FMB), S-cone ON bipolar (BB), and rod bipolar (RB). (Repro-

duced, with permission, from Boycott and Wässle 1999.)

撤光细胞

给光细胞

外丛状层

内核层

内丛状层

神经节

细胞层

弥散

双极

细胞 5

给光

侏儒

双极

细胞

弥散

双极

细胞 6

S视锥

给光

双极

细胞

弥散

双极

细胞 1

撤光

侏儒

双极

细胞

弥散

双极

细胞 2

弥散

双极

细胞 3

弥散

双极

细胞 4

excitatory signals from bipolar cells at glutamatergic

synapses. Some amacrine cells feed back directly to the

presynaptic bipolar cell at a reciprocal inhibitory synapse.

Some amacrine cells are electrically coupled to others

of the same type, forming an electrical network much

like that of the horizontal cells.

Through this inhibitory network, a bipolar cell ter-

minal can receive inhibition from distant bipolar cells,

in a manner closely analogous to the lateral inhibition of

photoreceptor terminals (see Figure 22–3A). Amacrine

cells also inhibit retinal ganglion cells directly. These

lateral inhibitory connections contribute substantially

to the antagonistic receptive field component of retinal

ganglion cells.

Temporal Filtering Occurs in Synapses and

Feedback Circuits

For many ganglion cells, a step change in light inten-

sity produces a transient response, an initial peak in

firing that declines to a smaller steady rate (see Figure

22–10). Part of this sensitivity originates in the negative-

feedback circuits involving horizontal and amacrine

cells. For example, a sudden decrease in light intensity

depolarizes the cone terminal, which excites the hori-

zontal cell, which in turn repolarizes the cone terminal

(see Figure 22–3A). Because this feedback loop involves

a brief delay, the voltage response of the cone peaks

abruptly and then settles to a smaller steady level. Simi-

lar processing occurs at the reciprocal synapses between

bipolar and amacrine cells in the inner retina.

In both cases, the delayed-inhibition circuit favors

rapidly changing inputs over slowly changing inputs.

The effects of this filtering, which can be observed

in visual perception, are most pronounced for large

stimuli that drive the horizontal and amacrine cell net-

works most effectively. For example, a large spot can

be seen easily when it flickers at a rate of 10 Hz but not

at a low rate (see Figure 22–14).

In addition to these circuit properties, certain cel-

lular processes contribute to shaping the temporal

response. For example, the AMPA-kainate type of

glutamate receptor undergoes strong desensitization.

A step increase in the concentration of glutamate at

the dendrite of a bipolar or ganglion cell leads to an

immediate opening of additional glutamate receptors.

As these receptors desensitize, the postsynaptic con-

ductance decreases again. The effect is to render a step

response more transient.

Retinal circuits seem to go to great lengths to speed

up their responses and emphasize temporal changes.

One likely reason is that the very first cell in the reti-

nal circuit, the photoreceptor, is exceptionally slow

(see Figure 22–7C). Following a flash of light, a cone

takes about 40 ms to reach the peak response, an intol-

erable delay for proper visual function. Through the

various filtering mechanisms in retinal circuitry, sub-

sequent neurons respond most vigorously during the

rising phase of the cone’s response. Indeed, some gan-

glion cells have a response peak only 20 ms after the

flash. Temporal processing in the retina clearly helps

to reduce visual reaction times, a life-extending trait as

Kandel-Ch22_0521-0544.indd 537 09/12/20 4:49 PM

视杆

双级

细胞

图 22.4.1: 猕猴视网膜中的双极细胞。细胞根据其在内网层中的末端分枝的深度排列。划分该层的远端（上）和

近端（下）水平的水平线表示撤光和给光型细胞的轴突末端之间的边界。假定上半部分的端子是撤光细胞的端

子，下半部分的端子是给光单元的端子。细胞类型是弥散双极细胞、给光和撤光小型双极细胞、短波锥双极细胞

和杆状双极细胞。

双极给光细胞和撤光细胞的形状不同，尤其是在轴突终止的内网层内的水平上。如图 22.4.1 所示，给光细

胞的轴突终止于近端（下）一半，而撤光细胞的轴突终止于远端（上）一半。在那里，它们在无长突细胞和神经

节细胞的树突上形成特定的突触连接。如图 22.0.3A 所示，给光双极激发给光神经节细胞，而撤光双极细胞激发

撤光神经节细胞。因此，视网膜输出的 2 个主要细分，给光和撤光通路，已经在双极细胞水平上建立。

如图 22.4.1 所示，双极细胞也可以通过其树突的形态来区分。在灵长类动物视网膜的中央区域，小型双极细

胞接收来自单个视锥细胞的输入并激发 P 型神经节细胞。这解释了为什么小细胞感受野的中心如此之小。弥漫

性双极细胞接收来自许多视锥细胞的输入并激发中波型神经节细胞。因此，大细胞的感受野中心要大得多。因

此，神经节细胞群中的刺激表征起源于专用的双极细胞通路，这些通路的区别在于它们与光受体和突触后目标

的选择性连接。

468

22.4 中间神经元网络塑造视网膜输出

22.4.2 空间滤波是通过侧抑制实现的

如图 22.0.3A 所示，并行开和关通路中的信号通过与水平细胞和无长突细胞的相互作用而改变。水平细胞具

有广泛的树枝状树突，这些树突在外丛状层中横向分布。光受体在与双极细胞共享的谷氨酸能末端接触这些棘

的尖端。此外，水平单元通过间隙连接彼此电耦合。

水平细胞有效地测量了广阔区域中光受体群体的平均激发水平。该信号通过抑制性突触反馈到感光器终端。

因此，感光器终端受到 2 种相反的影响：落在受体上的光使其超极化，但落在周围区域的光通过水平细胞的符

号反转突触使其去极化。因此，双极细胞具有拮抗的感受野结构。

感受野中的这种空间对抗通过内部视网膜中无长突细胞的侧向抑制得到增强。无长突细胞是神经元，其突

起仅在内网层中分支。大约有 30 种类型的无长突细胞是已知的，一些具有只有几十微米宽的小分枝，而另一些

则具有延伸到整个视网膜的突起。无长突细胞通常在谷氨酸能突触处接收来自双极细胞的兴奋信号。一些无长

突细胞在相互抑制性突触处直接反馈给突触前双极细胞。一些无长突细胞与其他相同类型的无长突细胞电耦合，

形成与水平细胞非常相似的电网络。

如图 22.0.3A 所示，通过这个抑制网络，双极细胞末端可以接受来自远处双极细胞的抑制，其方式与光受体

末端的侧抑制非常相似。无长突细胞也直接抑制视网膜神经节细胞。这些侧抑制连接对视网膜神经节细胞的拮

抗性感受野成分有很大贡献。

22.4.3 时间滤波发生在突触和反馈回路中

如图 22.3.1 所示，对于许多神经节细胞，光强度的阶跃变化会产生瞬态响应，放电的初始峰值会下降到较

小的稳定速率。这种敏感性部分源于涉及水平细胞和无长突细胞的负反馈回路。例如，如图 22.0.3A 所示，光强

度的突然降低使锥体末端去极化，从而激发水平细胞，进而使锥体末端重新极化。因为这个反馈回路涉及一个

短暂的延迟，锥体的电压响应突然达到峰值，然后稳定到一个较小的稳定水平。类似的过程发生在视网膜内部

双极细胞和无长突细胞之间的相互突触处。

对比灵敏度

100

对比灵敏度

100

2000 Td

200

频率（赫兹）

110

频率（赫兹）

110100

A 人类受试者 B 猕猴神经节细胞

100

图 22.4.2: 时间对比敏感度。A. 人类受试者对时间闪烁的敏感性是通过类似于图 22.5.3A 中的方法测量的，但刺

激是一个大点，其强度随时间呈正弦曲线变化，而不是在空间中。检测所需的阈值对比度的倒数相对于正弦闪

烁的频率绘制。灵敏度在高频和低频均下降。平均光照水平发生变化，从顶部到底部轨迹减少了 10 倍。B. 猕猴

视网膜中中波型神经节细胞的闪烁敏感性是通过 A 部分中应用于人类受试者的相同方法测量的。神经反应的检

测阈值定义为细胞放电中每秒 20 个脉冲的变化速率与闪烁同相。

在这 2 种情况下，延迟抑制回路有利于快速变化的输入而不是缓慢变化的输入。这种过滤的效果可以在视

觉感知中观察到，对于最有效地驱动水平和无长突细胞网络的大刺激最为明显。例如，如图 22.4.2 所示，当一个

大光点以 10 赫兹的频率闪烁时很容易看到，但以低频率闪烁时却看不到。

469

22.4 中间神经元网络塑造视网膜输出

除了这些回路特性外，某些细胞过程还有助于塑造时间响应。例如，α-氨基-3-羟基-5-甲基-4-异恶唑丙酸-红

藻氨酸类型的谷氨酸受体会经历强烈的脱敏作用。双极或神经节细胞树突处谷氨酸浓度的逐步增加导致额外的

谷氨酸受体立即打开。随着这些受体脱敏，突触后电导再次降低。效果是使阶跃响应更加瞬态。

视网膜回路似乎竭尽全力加快他们的反应并强调时间变化。如图 22.2.1C 所示，一个可能的原因是视网膜回

路中的第一个细胞，即光受体，异常缓慢。在一道闪光之后，视锥细胞需要大约 40 毫秒才能达到峰值响应，这

对于正常的视觉功能来说是无法忍受的延迟。通过视网膜回路中的各种过滤机制，后续神经元在视锥细胞反应

的上升阶段反应最强烈。事实上，一些神经节细胞在闪光后仅 20 毫秒就有一个响应峰值。视网膜中的时间处理

显然有助于减少视觉反应时间，这是一种延长寿命的特性，在高速公路交通中和在我们祖先的热带草原上一样

重要。

22.4.4 色觉始于锥形选择性回路

纵观有记载的历史，哲学家和科学家一直对色彩感知着迷。这种兴趣最初是由颜色与艺术的相关性驱动的，

后来是由它与光的物理特性关系驱动，最后是由电视和摄影中的商业利益驱动的。19 世纪出现了大量解释色彩

感知的理论，其中有 2 个在现代审查中幸存下来。它们基于仔细的心理物理学，对潜在的神经机制施加了强烈

的限制。

早期的实验表明，任何给定的自然光都可以通过将适量的 3 种原色光混合在一起来进行颜色匹配。这导致

了基于 3 种机制对光吸收的颜色感知三原色理论，每种机制具有不同的灵敏度光谱。如图 22.2.3 所示，这些对

应于 3 种锥体类型，其测量的吸收光谱充分解释了正常个体和色素基因遗传异常者的配色结果。

提出了所谓的对立过程理论来解释我们对不同色调的感知。根据这一理论，色觉涉及 3 个过程，它们对不

同颜色的光以相反的方式做出反应：（黄–蓝）会被黄色光刺激并被蓝光抑制；（红–绿）受红色刺激，受绿色抑

制；和（白–黑）受白色刺激，受黑色抑制。我们认识到视网膜后受体回路中的一些 19 世纪假设。

在人类视网膜的中央 10°，接收来自单个视锥细胞输入的单个小型双极细胞会激发每个 P 型神经节细胞。例

如，长波-给光神经节细胞的感受野中心由单个长波锥体和包含长波和中波锥体混合的拮抗环绕组成。当这个神

经元的感受野受到一个延伸到中心和周围的大而均匀的光点的刺激时，这个神经元被红光去极化并被绿光超极

化。类似的拮抗作用适用于其他 3 个小细胞：长波-撤光、中波-给光和中波-撤光。这些小细胞将它们的信号发送

到外侧膝状体核的小细胞层。

一种专用类型的短波-给光双极细胞选择性地收集短波锥体的信号，并将它们传输到小双层型神经节细胞。

因为这个神经节细胞也接受来自长波-撤光和中波-撤光双极细胞的激发，所以它被蓝光去极化并被黄光超极化。

另一种类型的神经节细胞显示出相反的特征：短波-撤光和（长波 + 中波）-给光。这些信号被传输到外侧膝状体

核的粒状细胞层。

大细胞被弥漫性双极细胞激发，这些细胞反过来收集来自许多视锥细胞的输入，无论色素类型如何。因此，

这些神经节细胞具有宽光谱灵敏度的大感受野。它们的轴突投射到外侧膝状体核的大细胞层。

以这种方式，彩色信号被视网膜组合和编码，以传输到丘脑和皮层。在初级视觉皮层的回路中，这些信号以

不同的方式重新组合，从而导致各种各样的感受野布局。只有大约 10% 的皮层神经元优先受颜色对比而不是亮

度对比的驱动。这可能反映了这样一个事实，即色觉（尽管它具有巨大的审美吸引力）对我们的整体健康只做

出了很小的贡献。作为这方面的一个例子，回想一下色盲个体，他们在某种意义上已经失去了一半的色彩空间，

可以在成长过程中从未注意到这一缺陷。

22.4.5 先天性色盲有多种形式

很少有人是真正的色盲，即完全无法区分颜色的变化和光强的变化，但许多人的色觉受损，并且在区分对

我们大多数人来说微不足道的东西时遇到困难，例如在红色和绿色之间。大多数此类色觉异常是先天性的，并

且已被详细描述；其他一些异常是由视觉通路的损伤或疾病引起的。

有些人只有两类视锥细胞而不是 3 类。这些二色视者发现很难或不可能区分某些颜色对三色视者来说明显

不同的表面。二色视者的问题是每个表面反射函数都由二值描述而不是三值描述表示，这种简化的描述导致二

470

22.4 中间神经元网络塑造视网膜输出

色视者比三色视者混淆更多的表面。如图 22.4.3 所示，色盲的简单测试利用了这一事实。

538 Part IV / Perception

Figure 22–16 A test for some forms of color blindness.The

numerals embedded in this color pattern can be distinguished

by people with trichromatic vision but not by dichromats who

are weak in red–green discrimination. If you don’t see any num-

bers please have your vision tested. (Reproduced, with permis-

sion, from Ishihara 1993.)

important in highway traffic as on the savannas of our

ancestors.

Color Vision Begins in Cone-Selective Circuits

Throughout recorded history, philosophers and sci-

entists have been fascinated by color perception. This

interest was originally driven by the relevance of color

to art, later by its relation to the physical properties of

light, and finally by commercial interests in television

and photography. The 19th century witnessed a profu-

sion of theories to explain color perception, of which

two have survived modern scrutiny. They are based on

careful psychophysics that placed strong constraints

on the underlying neural mechanisms.

Early experiments demonstrated that any given

natural light could be color-matched by mixing

together appropriate amounts of three primary lights.

This led to the trichromatic theory of color perception

based on absorption of light by three mechanisms,

each with a different sensitivity spectrum. These corre-

spond to the three cone types (see Figure 22–6), whose

measured absorption spectra fully explain the color-

matching results both in normal individuals and those

with genetic anomalies in the pigment genes.

The so-called opponent-process theory was pro-

posed to explain our perception of different hues. Accord-

ing to this theory, color vision involves three processes

that respond in opposite ways to light of different colors:

(y–b) would be stimulated by yellow and inhibited by

blue light; (r–g) stimulated by red and inhibited by green;

and (w–bk) stimulated by white and inhibited by black.

We recognize some of these 19th century postulates in the

postreceptor circuitry of the retina.

In the central 10° of the human retina, a single

midget bipolar cell that receives input from a single

cone excites each P-type ganglion cell. An L-ON gan-

glion cell, for example, has a receptive field center con-

sisting of a single L cone and an antagonistic surround

involving a mixture of L and M cones. When this neu-

ron’s receptive field is stimulated with a large uniform

spot of light that extends over both the center and the

surround, this neuron is depolarized by red light and

hyperpolarized by green light. Similar antagonism

holds for the three other P-cells: L-OFF, M-ON, and

M-OFF. These P-cells send their signals to the parvo-

cellular layers of the lateral geniculate nucleus.

A dedicated type of S-ON bipolar cell collects the

signals of S-cones selectively and transmits them to

ganglion cells of the small bistratified type. Because

this ganglion cell also receives excitation from L-OFF

and M-OFF bipolar cells, it is depolarized by blue light

and hyperpolarized by yellow light. Another type of

ganglion cell shows the opposite signature: S-OFF and

(L + M)-ON. These signals are transmitted to the koni-

ocellular layers of the lateral geniculate nucleus.

The M-cells are excited by diffuse bipolar cells,

which in turn collect inputs from many cones regard-

less of pigment type. These ganglion cells therefore

have large receptive fields with broad spectral sensi-

tivity. Their axons project to the magnocellular layers

of the lateral geniculate nucleus.

In this way, chromatic signals are combined and

encoded by the retina for transmission to the thalamus

and cortex. In circuits of the primary visual cortex, these

signals are recombined in different ways, leading to a

great variety of receptive field layouts. Only about 10%

of cortical neurons are preferentially driven by color con-

trast rather than luminance contrast. This likely reflects

the fact that color vision—despite its great aesthetic

appeal—makes only a small contribution to our overall

fitness. As an illustration of this, recall that colorblind

individuals, who in a sense have lost half of their color

space, can grow up without ever noticing that defect.

Congenital Color Blindness Takes Several Forms

Few people are truly colorblind in the sense of being

wholly unable to distinguish a change in color from a

Kandel-Ch22_0521-0544.indd 538 09/12/20 4:49 PM

图 22.4.3: 对某些形式的色盲的测试。嵌入这种颜色图案中的数字可以被具有三色视觉的人辨别出来，但不能被

红绿辨别力较弱的二色视者辨别出来。如果您没有看到任何数字，请检查您的视力。

虽然存在 3 种形式的双色性，对应于 3 种类型视锥细胞中每一种的损失，但有 2 种比第三种更常见。常见

形式对应于长波视锥细胞或中波视锥细胞的丧失，分别称为红色盲和绿色盲。红色盲和绿色盲几乎总是发生在

男性身上，各占 1% 左右。这些病症由自身未受影响的女性传播，因此与 X 染色体上的基因有关。第三种形式

的双色性，蓝色盲，涉及短波锥体的丧失或功能障碍。它仅影响大约 1 万人中的 1 人，以相同的频率折磨女性和

男性，并且涉及 7 号染色体上的一个基因。

由于长波和中波视锥细胞大量存在，人们可能会认为其中一种或另一种视锥细胞的缺失会对视力造成更广

泛的损害，而不仅仅是削弱色觉。事实上，这不会发生，因为二色视视网膜中长波和中波视锥细胞的总数没有改

变。所有注定要成为长波或中波视锥细胞的细胞可能在绿色盲中转化为长波视锥细胞，在红色盲中转化为中波

视锥细胞。

除了以二色性为代表的相对严重的色盲形式外，还有较温和的形式，同样主要影响男性。这些所谓的异常

三色视者的视锥细胞的光谱灵敏度与正常三色视者不同。异常三色性是由于一种正常的视锥细胞色素被具有不

同光谱灵敏度的改变蛋白质所取代所致。2 种常见形式，红色弱和绿色弱，共同影响大约 7% 的男性，分别代表

长波或中波视锥细胞被具有某种中间光谱敏感性的色素取代。

色觉缺陷的遗传学已广为人知。如图 22.4.4A 所示，长波和中波色素的基因以头尾排列的形式位于 X 染色

体上。色素蛋白具有非常相似的结构，仅 4% 的氨基酸不同。具有正常色觉的人拥有长波色素基因的单个拷贝，

以及 1 到 3 个（有时多达 5 个），几乎相同的中波色素基因拷贝。

如图 22.4.4B 所示，这些基因的接近性和相似性使它们易于发生各种形式的重组，导致基因丢失或形成杂交

基因，从而导致常见的红-绿缺陷形式。对二色视者中这些基因的检查揭示了红色盲中长波色素基因的缺失和绿

色盲中一种或多种中波色素基因的缺失。异常三色视者具有长波-中波或中波-长波混合基因，这些基因编码具有

偏移光谱灵敏度的视觉色素；转变的程度取决于重组点。在蓝色盲中，短波锥体功能的丧失是由短波色素基因

突变引起的。

22.4.6 杆状和锥状回路在视网膜内部合并

如图 22.0.3B 所示，对于弱光条件下的视觉，哺乳动物视网膜有一个给光双极细胞，专门连接到视杆。通过

收集来自多达 50 个视杆的输入，这种视杆双极细胞可以将分散的单光子吸收效应集中在一小块视网膜中。没有

471

22.4 中间神经元网络塑造视网膜输出

Chapter 22 / Low-Level Visual Processing: The Retina 539

Figure 22–17 L- and M-pigment genes on the X

chromosome.

A.The L- and M-pigment genes normally lie next to each other

on the chromosome. The base of each arrow corresponds to

the 5′ end of the gene, and the tip corresponds to the 3′ end.

Males with normal color vision can have one, two, or three

copies of the gene for the M pigment on each X chromosome.

(Adapted, with permission, from Nathans, Thomas, and

B.Recombinations of the L- and M-pigment genes can lead to

the generation of a hybrid gene (3 and 4) or the loss of a gene

(1), the patterns observed in colorblind men. Spurious recombi-

nation can also cause gene duplication (2), a pattern observed

in some people with normal color vision. (Adapted from Streyer

1988. Used with permission from J. Nathans.)

change in the intensity of light, but many individuals

have impaired color vision and experience difficulties

in making distinctions that for most of us are trivial,

for example between red and green. Most such abnor-

malities of color vision are congenital and have been

characterized in detail; some other abnormalities result

from injury or disease of the visual pathway.

Some people have only two classes of cones instead

of three. These dichromats find it difficult or impossi-

ble to distinguish some surfaces whose colors appear

distinct to trichromats. The dichromat’s problem is that

every surface reflectance function is represented by a

two-value description rather than a three-value one,

and this reduced description causes dichromats to con-

fuse many more surfaces than do trichromats. Simple

tests for color blindness exploit this fact (Figure 22–16).

Although there are three forms of dichromacy, cor-

responding to the loss of each of the three types of cones,

two kinds are much more common than the third. The

common forms correspond to the loss of the L cones

or M cones and are called protanopia and deuteranopia,

respectively. Protanopia and deuteranopia almost

always occur in males, each with a frequency of about

1%. The conditions are transmitted by women who are

not themselves affected, and so implicate genes on the

X chromosome. A third form of dichromacy, tritanopia,

involves loss or dysfunction of the S cone. It affects

only about 1 in 10,000 people, afflicts women and men

with equal frequency, and involves a gene on chromo-

some 7.

Because the L and M cones exist in large numbers,

one might think that the loss of one or the other type

would impair vision more broadly than just weaken-

ing color vision. In fact, this does not happen because

the total number of L and M cones in the dichromat

retina is not altered. All cells destined to become L or

M cones are probably converted to L cones in deuter-

anopes and to M cones in protanopes.

In addition to the relatively severe forms of color-

blindness represented by dichromacy, there are milder

forms, again affecting mostly males. These so-called

anomalous trichromats have cones whose spectral

sensitivities differ from those in normal trichromats.

Anomalous trichromacy results from the replacement

of one of the normal cone pigments by an altered pro-

tein with a different spectral sensitivity. Two common

forms, protanomaly and deuteranomaly, together

affect about 7% of males and represent, respectively,

the replacement of the L or M cones by a pigment with

some intermediate spectral sensitivity.

The genetics of color vision defects are well under-

stood. The genes for the L and M pigments reside on

the X chromosome in a head-to-tail arrangement

(Figure 22–17A). The pigment proteins have very simi-

lar structures, differing in only 4% of their amino acids.

People with normal color vision possess a single copy

of the gene for the L pigment and from one to three—

occasionally as many as five—nearly identical copies

of the gene for the M pigment.

The proximity and similarity of these genes predis-

poses them to varied forms of recombination, leading

either to the loss of a gene or to the formation of hybrid

genes that account for the common forms of red–green

defect (Figure 22–17B). Examination of these genes in

dichromats reveals a loss of the L-pigment gene in pro-

tanopes and a loss of one or more M-pigment genes

in deuteranopes. Anomalous trichromats have L-M

or M-L hybrid genes that code for visual pigments

with shifted spectral sensitivity; the extent of the shift

长波-色素中波-色素

Kandel-Ch22_0521-0544.indd 539 09/12/20 4:49 PM

图 22.4.4: X 染色体上的长波和中波色素基因。A. 长波-和中波-色素基因通常在染色体上彼此相邻。每个箭头的

底部对应基因的 5’ 端，尖端对应基因的 3’ 端。具有正常色觉的男性在每条 X 染色体上可能有 1 个、2 个或 3 个

中波色素基因拷贝

[185]

。B. 长波-和中波-色素基因的重组可导致杂合基因（3 和 4）的产生或丢失基因（1）的模

式，在色盲男性中观察到的模式。虚假重组也可能导致基因复制（2），这是在一些具有正常色觉的人身上观察

到的模式。

472

22.5 视网膜的灵敏度适应光照的变化

相应的专用于杆的撤光双极细胞。

与所有其他双极细胞不同，杆状双极细胞不直接接触神经节细胞，而是激发专用神经元，即 AII 无长突细

胞。该无长突细胞接收来自多个杆状双极细胞的输入，并将其输出传送至锥状双极细胞。它通过间隙连接向给

光双极细胞提供兴奋信号，并向撤光双极提供甘氨酸能抑制信号。如前所述，这些锥形双极细胞依次激发给光

和撤光神经节细胞。因此，棒状信号在绕道后被馈入锥体系统，为开和关路径产生适当的信号极性。添加的中间

神经元的目的可能是允许更多的杆状信号汇集而不是锥状信号。

杆信号也通过另外两条通路进入锥体系统。杆可以直接通过电接头驱动相邻的锥体，并且它们与主要为锥

体服务的撤光双极细胞连接。一旦视杆细胞信号通过这些通路到达视锥细胞双极，它就可以利用内部视网膜同

样复杂的回路。因此，哺乳动物视网膜的视杆系统可能是进化后添加到视锥回路中的想法。

22.5 视网膜的灵敏度适应光照的变化

视觉在许多不同的照明条件下运行。来自物体的光的强度取决于环境照明的强度和物体表面反射的光的比

例，称为反射率。一天中遇到的强度范围很大，变化跨越 10 个数量级，但这种变化的大部分对于指导行为是无

用的。

Chapter 22 / Low-Level Visual Processing: The Retina 541

Figure 22–18 A brightness illusion.

A.The two tiles marked with small dots appear to have differ-

ent color but actually reflect the same light intensity. (To see

this, fold the page so they touch.) The trace underneath plots

a profile of light intensity at the level of the arrowheads. Your

visual system interprets this retinal image as a regular tile pat-

tern under spatially varying illumination with a diffuse shadow

in the right half. Under that interpretation, the right tile must

have a lighter color than the left, which is what you perceive.

This process is automatic and requires no conscious analysis.

B.Retinal processing contributes to the perception of “light-

ness” by discounting the shadow’s smooth gradients of

illumination and accentuating the sharp edges between

checkerboard fields. The receptive field for a visual neuron with

an excitatory center and inhibitory surround is shown at the

top. As shown in a hundredfold magnification at the bottom,

the surround is weak but extends over a much larger area than

the center.

C.The result when a population of visual neurons with

receptive fields as in B processes the image in A. This opera-

tion—the convolution of the image in A with the profile in

B—subtracts from each point in the visual field the average

intensity in a large surrounding region. The neural representa-

tion of the object has largely lost the effects of shading, and

the two tiles in question do indeed have different brightness

values in this representation.

x100

ABC

be proportional to the background intensity, a rela-

tionship known as Weber’s law of adaptation, which we

encountered in considering somatic receptor sensitiv-

ity (Chapter 17). The visual system follows Weber’s law

approximately: Over the entire range of vision, sensi-

tivity decreases somewhat less steeply with increasing

background intensity (Figure 22–19B).

Multiple Gain Controls Occur Within the Retina

The enormous change in gain required for light adap-

tation arises at multiple sites within the retina. In star-

light, a single rod cell is stimulated by a photon only

every few seconds, a rate insufficient to alter the cell’s

adaptation status. However, a retinal ganglion cell

combines signals from many rods, thus receiving a

steady stream of photon signals that can elicit a light-

dependent gain change in the cell.

At somewhat higher light intensities, a rod bipo-

lar cell begins to adapt, changing its responsiveness

depending on the average light level. Next, we reach

a light intensity at which the gain of individual rod

cells gradually decreases. Beyond that, the rods satu-

rate: All their cGMP-dependent channels are closed,

and the membrane potential no longer responds to the

light stimulus. By this time, around dawn, the much

less sensitive cone cells are being stimulated effectively

and gradually take over from the rods. As the ambient

light increases further, toward noon, light adaptation

results principally from gain changes within the cones.

The cellular mechanisms of light adaptation are

best understood in the photoreceptors. The calcium-

dependent feedback pathways discussed earlier have

a prominent role. Recall that when a light flash closes

the cGMP-gated channels, the resulting decrease

in intracellular Ca

accelerates several biochemi-

cal reactions that terminate the response to the flash

(see Figure 22–7B). When illumination is continuous,

however, the Ca

concentration remains low, and

all these reactions are therefore in a steady state that

both lowers the gain and accelerates the time course

of the receptor’s response to light (Figure 22–19C). As

a result, the light-adapted photoreceptor can respond

to rapid changes in intensity much more quickly. This

has important consequences for human visual percep-

tion; the contrast sensitivity to high-frequency flicker

increases with intensity, an effect observed in primate

retinal ganglion cells as well (see Figure 22–14).

Kandel-Ch22_0521-0544.indd 541 09/12/20 4:49 PM

图 22.5.1: 亮度错觉。A. 标有小点的 2 块瓷砖看起来颜色不同，但实际上反射的光强度相同（要看到这一点，请

折叠页面，使它们接触）。下方的迹线绘制了箭头水平的光强度分布图。您的视觉系统将此视网膜图像解释为空

间变化照明下的规则图块图案，右半部分有漫射阴影。根据这种解释，右边的瓷砖必须比左边的颜色浅，这就

是您所感知的。这个过程是自动的，不需要有意识的分析。B. 视网膜处理通过降低阴影平滑的照明梯度并突出

棋盘区域之间的锐利边缘来促进“亮度”的感知。具有兴奋中心和抑制周围的视觉神经元的感受野显示在顶部。

如底部放大百倍所示，环绕声较弱，但延伸的区域比中心区域大得多。C. 当一群具有 B 中感受野的视觉神经元

处理 A 中的图像时的结果。这个操作（A 中的图像与 B 中的轮廓的卷积）从视野中的每个点减去 A 中的平均强

度一个大的周边地区。目标的神经表示在很大程度上失去了阴影的影响，并且所讨论的 2 个图块在该表示中确

实具有不同的亮度值。

照明强度变化约 9 个数量级，主要是因为我们的星球每天围绕其轴旋转一次，而物体反射率变化小得多，在

典型场景中变化约一个数量级。但是这种反射率对于视觉来说是一个有趣的量，因为它表征了物体并将它们与

背景区分开来。事实上，如图 22.5.1 所示，我们的视觉系统非常擅长独立于环境光照计算表面反射率。

随着环境照明的整体增加，视觉场景中的所有点都以相同的因子变亮。如果眼睛可以简单地将其灵敏度降

低相同的因素，则图像的神经表征将在神经节细胞水平上保持不变，并且可以由大脑的其余部分以与光照变化

之前相同的方式进行处理。此外，由于不同的物体反射率，视网膜神经节细胞只需要编码 10 倍范围的图像强度，

而不是包括环境照明变化的 100 亿倍范围。部分灵敏度调整是由瞳孔执行的，瞳孔在强光下收缩，将视网膜照

明度降低多达 10 倍。此外，视网膜本身执行自动增益控制，称为光适应，接近理想的正常化我们在这里想象过。

473

22.5 视网膜的灵敏度适应光照的变化

22.5.1 光适应在视网膜处理和视觉感知中很明显

如图 22.5.2A 所示，当不同强度的闪光呈现恒定的背景照明时，视网膜神经节细胞的反应符合 S 形曲线。最

弱的闪光不会引起反应，闪光强度的分级增加会引起分级反应，最亮的闪光会引起饱和。当背景照明增加时，响

应曲线保持相同的形状，但转向更高的闪光强度。为了补偿背景照明的增加，神经节细胞现在对光变化不太敏

感：在较高背景下，需要更大的变化才能引起相同的反应。这种刺激-反应关系的横向移动是视网膜光适应的标

志。

542 Part IV / Perception

放电率

(脉冲/秒)

300

毫伏 (mV/Rh*)

–0.5

–1.0

归一化响应

0.0

–0.5

–1.0

0.00.5

时间（秒）

1.01.5 2.0

9⋅10

–5

9⋅10

–4

9⋅10

–3

9⋅10

–2

9⋅10

–1

对数（阈值闪光）

–3

–4

–5

–6

–7

–8

–1–2–3–4–5–6–7–8–9

(

零

)

对数

（背景）

–5 –4 –3 –2 –1 0

200

100

猫的神经节细胞

人类受试者

测试点亮度

(cd/m

)

韦伯定理斜坡

视杆

视锥

对数（背景）

星光

黄昏中午

0.0

Figure 22–19 Light adaptation.

A.The receptive field of a cat retinal ganglion cell was illumi-

nated uniformly at a steady background intensity, and a test

spot was flashed briefly on the receptive field center. The peak

firing rate following the flash was measured and plotted against

the logarithm of the flash intensity. Each curve corresponds

to a different background intensity, increasing by factors of 10

from left to right. (Reproduced, with permission, from Sakmann

B.A small test spot was flashed briefly on a steadily illumi-

nated background, and the flash intensity gradually increased to

where a human subject could just detect it. The procedure was

repeated at different background intensities. Here, the thresh-

old flash intensity is plotted against the background intensity.

The curve has two branches connected by a distinct kink: These

correspond to the regimes of rod and cone vision. The slope

of Weber’s law represents the idealization when the threshold

intensity is proportional to the background intensity. (Adapted

from Wyszecki and Stiles 1982.)

C.The top plot shows the responses of a macaque monkey’s

rod cell to flashes presented at varying background intensi-

ties. The cell’s single-photon response was calculated from

the recorded membrane potential divided by the number of

rhodopsins (Rh) activated by the flash. The gain of the single-

photon response decreases substantially with increasing back-

ground intensity. The background intensity, in photon/μm

/s, is

0 for trace 0, 3.1 for trace 1, 12 for trace 2, 41 for trace 3, 84 for

trace 4, and 162 for trace 5. In the bottom plot, the same data

(except for the smallest response) are normalized to the same

amplitude, showing that the time course of the single-photon

response accelerates at high intensity. (Reproduced, with per-

mission, from Schneeweis and Schnapf 2000.)

Kandel-Ch22_0521-0544.indd 542 09/12/20 4:49 PM

A 猕猴视杆细胞

图 22.5.2: 光照适应。A. 猫视网膜神经节细胞的感受野在稳定的背景强度下均匀照明，并在感受野中心短暂闪烁

一个测试点。测量闪光后的峰值激活率并根据闪光强度的对数绘制。每条曲线对应不同的背景强度，从左到右

增加 10 倍

[186]

。B. 一个小的测试点在稳定照明的背景上短暂闪光，闪光强度逐渐增加到人类受试者刚好可以检

测到的程度。在不同的背景强度下重复该过程。这里，阈值闪光强度是根据背景强度绘制的。该曲线有 2 个分

支，由一个明显的扭结相连：它们对应于视杆和视锥视觉的状态。当阈值强度与背景强度成正比时，韦伯定律的

斜率表示理想化

[187]

。C. 上图显示了猕猴视杆细胞对不同背景强度下出现的闪光的反应。细胞的单光子响应是根

据记录的膜电位除以闪光激活的视紫红质的数量计算得出的。单光子响应的增益随着背景强度的增加而显著降

低。背景强度（以光子/平方微米/秒为单位）对于迹线 0 为 0，对于迹线 1 为 3.1，对于迹线 2 为 12，对于迹线 3

为 41，对于迹线 4 为 84，对于迹线 5 为 162。在底部图中，相同数据（最小响应除外）被归一化为相同的振幅，

表明单光子响应的时间进程在高强度下加速

[188]

。

这种增益变化对人类视觉感知的影响在心理物理学实验中显而易见。如图 22.5.2B 所示，当人类受试者被要

求在恒定照明的背景场中检测闪光时，在更亮的背景上检测需要更亮的闪光。在前面讨论的理想增益控制机制

474

22.6 亮点

下，如果 2 个刺激引起背景强度的相同分数变化，它们将产生相同的响应。在那种情况下，阈值闪光强度应与背

景强度成正比，这种关系称为韦伯适应定律，我们在考虑体细胞受体敏感性时遇到过这种关系（第 17 章）。如

图 22.5.2B 所示，视觉系统大致遵循韦伯定律：在整个视觉范围内，随着背景强度的增加，灵敏度的下降幅度有

所缓和。

22.5.2 多重增益控制发生在视网膜内

光适应所需增益的巨大变化出现在视网膜内的多个部位。在星光下，单个视杆细胞每隔几秒就会受到一个

光子的刺激，这个速度不足以改变细胞的适应状态。然而，视网膜神经节细胞结合来自许多视杆细胞的信号，从

而接收稳定的光子信号流，从而在细胞中引起光依赖性增益变化。

在稍高的光照强度下，杆状双极细胞开始适应，根据平均光照水平改变其响应能力。接下来，我们达到单个

杆细胞的增益逐渐降低的光强度。除此之外，视杆细胞饱和：它们所有依赖于环鸟苷-3,5-单磷酸盐的通道都关闭

了，膜电位不再对光刺激做出反应。到这个时候，大约在黎明时分，敏感度低得多的视锥细胞得到有效刺激，并

逐渐取代视杆细胞。随着环境光进一步增加，接近中午，光适应主要来自视锥细胞内的增益变化。

光受体的细胞机制在光受体中得到了最好的理解。前面讨论的钙依赖性反馈通路具有突出的作用。回想一

下，如图 22.2.1B 所示，当闪光关闭环鸟苷-3,5-单磷酸盐门控通道时，细胞内 Ca

的减少会加速几种生化反应，

从而终止对闪光的反应。然而，如图 22.5.2C 所示，当光照持续时，Ca

浓度仍然很低，因此所有这些反应都处

于稳定状态，既降低了增益又加速了受体对光的反应的时间进程。因此，光适应感光器可以更快地响应强度的

快速变化。这对人类视觉感知有重要影响；如图 22.4.2 所示，对高频闪烁的对比敏感度随着强度的增加而增加，

这也是在灵长类动物视网膜神经节细胞中观察到的效果。

22.5.3 光适应改变空间处理

除了视网膜反应的灵敏度和速度之外，光适应还改变了空间处理的规则。如图 22.3.1 所示，在明亮的光线下，

许多神经节细胞在它们的感受野中有一个尖锐的中心环绕结构。随着光线变暗，对立的环绕声变得宽广而微弱，

最终消失。在这些条件下，视网膜的回路只是简单地积累稀有光子，而不是计算局部强度梯度。如图 22.0.3 所

示，感受野特性的这些变化是由于水平和无长突细胞网络产生的侧抑制的变化而发生的。这些过程的一个重要

调节剂是多巴胺，由专门的无长突细胞以依赖光的方式释放。

这些视网膜效应在人类感知上留下了印记。在明亮的光线下，我们的视觉系统更喜欢精细的光栅而不是粗

糙的光栅。如文本框 22.1 和图 22.5.3 所示，但在昏暗的光线下，我们对粗光栅最敏感：随着中心环绕拮抗作用

的丧失，低空间频率不再衰减。

总之，光照适应有 2 个重要作用。一种是丢弃有关环境光强度的信息，同时保留有关物体反射率的信息。另

一个是将视网膜神经节细胞中的小动态放电范围与环境中的大范围光强度相匹配。这些大的增益变化必须在视

神经纤维中产生动作电位之前通过分级神经元信号来完成，因为这些纤维的放电率只能在 2 个数量级内有效变

化。事实上，光适应的关键需求可能就是为什么这种神经回路位于眼睛而不是视神经另一端的大脑中。

22.6 亮点

1. 视网膜将投射到光受体上的光模式转化为神经信号，这些信号通过视神经传送到大脑中专门的视觉中心。

不同的神经节细胞群沿着并行通路传输视网膜图像的多种神经表征。

2. 视网膜丢弃了受体水平上可用的大部分刺激信息，并提取了对中央视觉系统有用的视野的某些低层特征。

精细的空间分辨率仅在凝视中心的狭窄区域中保持。图像中的强度梯度，例如物体边缘，在空间均匀的部分上

得到强调；场景的不变部分增强了时间变化。

3. 视网膜能灵活适应视觉条件的变化，尤其是昼夜光照变化较大。关于绝对光照水平的信息大部分被丢弃，

有利于随后对场景内物体反射率的分析。

475

22.6 亮点

Figure 22–13 Spatial contrast sensitivity.

A.1. The contrast sensitivity of human subjects was

measured using gratings with different spatial frequen-

cies (see Figure 22–12). At each frequency, the contrast

was increased to the threshold for detection, and the

inverse of that contrast value was plotted against spatial

frequency, as shown here. The curves were obtained at

different mean intensities, decreasing by factors of 10

from the top to the bottom curve. (Reproduced, with

permission, from De Valois, Morgan, and Snodderly

1974.) 2. Contrast sensitivity of a P-type ganglion cell in

the macaque retina measured at high intensity. At each

spatial frequency, the contrast was gradually increased

until it produced a detectable change in the neuron’s

firing rate. The inverse of that threshold contrast was

plotted as in part A-1. The isolated dot at left marks the

sensitivity at zero spatial frequency, a spatially uniform

field. (Reproduced, with permission, from Derrington and

Lennie 1984.)

B.Stimulation of a center-surround receptive field with

sinusoid gratings. The neuron’s sensitivity to light at dif-

ferent points on the retina is modeled as a “difference-of-

Gaussians” receptive field, with a narrow positive Gaussian

for the excitatory center and a broad negative Gaussian for

the inhibitory surround. Multiplying the profile of the grating

stimulus (intensity vs position) with the profile of the recep-

tive field (sensitivity vs position) and integrating over all

space calculates the stimulus strength delivered by a partic-

ular grating. The resulting sensitivity of the receptive field to

gratings of different frequency is shown in the plot on the

right. At low spatial frequencies, the negative contribution

from the surround cancels the contribution from the center,

leading to a drop in the difference curve. (Reproduced, with

permission, from Enroth-Cugell and Robson 1984.)

Figure 22–14 Temporal contrast sensitivity.(Repro-

duced, with permission, from Lee et al. 1990.)

A.The sensitivity of human subjects to temporal flicker

was measured by methods similar to those in Figure

22–13A, but the stimulus was a large spot whose intensity

varied sinusoidally in time rather than in space. The inverse

of the threshold contrast required for detection is plotted

against the frequency of the sinusoidal flicker. Sensitivity

declines at both high and low frequencies. The mean light

level varied, decreasing by factors of 10 from the top to

the bottom trace.

B.The flicker sensitivity of M-type ganglion cells in the

macaque retina was measured by the same method applied

to human subjects in part A. The detection threshold for the

neural response was defined as a variation of 20 spikes per

second in the cell’s firing rate in phase with the flicker.

对比敏感度

响应率

1.0

0.1

0.01

0.5 1510 20 50

暗

A 人类和猴子的敏感性 B 神经节细胞感受野的敏感性

1 人类受试者

200

亮

100

低

中

高

中心

差别

(感受野概述)

周围

空间频率

0.01

0.1

空间频率（周期/度）

中心

低中高

差别

周围

对比敏感度

2 猴子神经节细胞

20100.1

空间频率（周期/度）

Contrast sensitivity

100

Contrast sensitivity

100

2000 Td

200

Frequency (Hz)

110

Frequency (Hz)

110100

A Human subjects B Macaque ganglion cells

100

Kandel-Ch22_0521-0544.indd 535 09/12/20 4:49 PM

图 22.5.3: 空间对比敏感度。A. 1. 人类受试者的对比敏感度是使用具有不同空间频率的光栅测量的（见图 22.5.4）。

在每个频率下，对比度增加到检测阈值，并且该对比度值的倒数相对于空间频率绘制，如图所示。这些曲线是在

不同的平均强度下获得的，从顶部到底部曲线减少了 10 倍

[189]

。2. 在高强度下测量的猕猴视网膜中 P 型神经节

细胞的对比敏感度。在每个空间频率下，对比度逐渐增加，直到它在神经元的放电率中产生可检测的变化。如

A-1 部分所示，绘制了该阈值对比度的倒数。左侧的孤立点标记了零空间频率下的灵敏度，这是一个空间均匀的

场

[190]

。B. 用正弦光栅刺激中心环绕感受野。神经元对视网膜上不同点的光的敏感性被建模为“高斯差分”感受

野，其中兴奋性中心为窄正高斯分布，抑制性周围为宽负高斯分布。将光栅刺激的轮廓（强度与位置）与感受野

的轮廓（灵敏度与位置）相乘，并在所有空间上积分，计算出特定光栅传递的刺激强度。右图显示了感受野对不

同频率光栅的灵敏度。在低空间频率下，环绕声的负贡献抵消了中心的贡献，导致差异曲线下降

[191]

。

Figure 22–12 Sinusoid grating displays used in psy-

chophysical experiments with human subjects.Such

stimuli were used in the experiments discussed in

Figure 22–13.

Although small spots of light are useful for probing the

receptive fields of single neurons in visual pathways,

different stimuli are needed to learn about human visual

perception. Grating stimuli are commonly used to probe

how our visual system deals with spatial and temporal

patterns.

The subject views a display in which the intensity

varies about the mean as a sinusoidal function of space

(Figure 22–12). Then the contrast of the display—defined

as the peak-to-peak amplitude of the sinusoid divided

by the mean—is reduced to a threshold at which the

grating is barely visible. This measurement is repeated

for gratings of different spatial frequencies.

When the inverse of this threshold is plotted against

the spatial frequency, the resulting contrast sensitivity

curve provides a measure of sensitivity of visual percep-

tion to patterns of different scales (Figure 22–13A). When

measured at high light intensity, sensitivity declines

sharply at high spatial frequencies, with an absolute

threshold at approximately 50 cycles per degree. This

sensitivity is limited fundamentally by the quality of the

optical image and the spacing of cone cells in the fovea

(see Figure 22–1C).

Interestingly, sensitivity also declines at low spatial

frequencies. Patterns with a frequency of approximately

5 cycles per degree are most visible. The visual system is

said to have band-pass behavior because it rejects all but

a band of spatial frequencies.

One can use the same techniques to measure the

sensitivity of individual retinal ganglion cells in pri-

mates. The results resemble those for human subjects

(Figure 22–13), suggesting that these basic features of

visual perception are determined by the retina.

The band-pass behavior can be understood on the

basis of spatial antagonism in center-surround recep-

tive fields. A very fine grating presents many dark and

bright stripes within the receptive-field center; their

effects cancel one another and thus provide no net exci-

tation. A very coarse grating presents a single stripe to

both the center and surround of the receptive field, and

their antagonism again provides the ganglion cell little

net excitation. The strongest response is produced by a

grating of intermediate spatial frequency that just covers

the center with one stripe and most of the surround with

stripes of the opposite polarity (Figure 22–13B).

In dim light, the visual system’s contrast sensitiv-

ity declines, but more so at high than at low spatial

frequencies (Figure 22–13A). Thus, the peak sensitivity

shifts to lower spatial frequencies, and eventually the

curve loses its peak altogether. In this state, the visual

system has so-called low-pass behavior, for it preferen-

tially encodes stimuli of low spatial frequency. The fact

that in dim light the receptive fields of ganglion cells

lose their antagonistic surrounds explains the transi-

tion from band-pass to low-pass spatial filtering

(Figure 22–13B).

Similar experiments can be done to test visual sen-

sitivity to temporal patterns. Here, the intensity of a test

stimulus flickers sinusoidally in time, while the contrast

is gradually brought to the threshold level of detection.

For humans, contrast sensitivity declines sharply at very

high flicker frequencies, but declines also at very low

frequencies (Figure 22–14A). Flicker at approximately

10 Hz is the most effective stimulus. One finds similar

band-pass behavior in the flicker sensitivity of macaque

retinal ganglion cells (Figure 22–14B).

Sensitivity to temporal contrast also depends on

the mean light level. For human subjects, the optimum

flicker frequency shifts downward at lower stimulus

intensities and the peak in the curve becomes less and

less prominent (Figure 22–14). The fact that primate reti-

nal ganglion cells duplicate this behavior suggests that

retinal processing limits the performance of the entire

visual system in these simple tasks.

Box 22–1 Spatiotemporal Sensitivity of Human Perception

低空间频率

高空间频率、高对比度

高空间频率、低对比度

Kandel-Ch22_0521-0544.indd 534 09/12/20 4:49 PM

图 22.5.4: 正弦波光栅显示器用于人类受试者的心理物理学实验。在图 22.5.3 中讨论的实验中使用了此类刺激。

476

22.6 亮点

4. 当色素分子吸收光子时，光刺激的转导开始于感光细胞的外段。这启动了一个放大的 G 蛋白级联反应，

最终降低了膜电导，使光受体超极化，并减少了突触处的谷氨酸释放。多种反馈机制，其中细胞内 Ca

具有重

要作用，用于关闭级联中的酶并终止光响应。

5. 杆状感光器是高效的光收集器并服务于夜间视觉。视锥细胞的敏感性要低得多，并且全天都在发挥作用。

视锥突触到双极细胞上，进而激发神经节细胞。视杆连接到专门的视杆双极细胞，其信号通过无长突细胞传递

到视锥双极细胞。

6. 垂直兴奋通路由主要是抑制性的水平连接调节。通过这些横向网络，神经节细胞感受野周围的光抵消了

中心光的影响。同样的负反馈回路也加强了神经节细胞的瞬态反应。

7. 将信息隔离到并行通路中以及通过抑制性横向连接形成反应特性是视觉系统中普遍存在的组织原则。

477